The European Securities and Markets Authority shows in its 2025 report that the volume of data used for monitoring financial markets is growing rapidly, and the quality of reporting has improved in several important areas. ESMA has started to use generative artificial intelligence in operational activities, including for supervisory analysis and market abuse detection projects, while working on simplifying reporting for firms.
The European Securities and Markets Authority increasingly uses data, automation, and artificial intelligence to monitor financial markets in the European Union. The ESMA report for 2025 shows that authorities receive billions of transactions, reports on financial instruments, funds, derivatives, securitizations, missing sales, digital incidents, and rating agencies, and the quality of this data determines how quickly risks to investors and markets can be detected.
In short ESMA publishes the sixth edition of the report on the quality and use of data. In 2025, the volume of transactions reported under MiFIR reached approximately 9.2 billion. The quality of data reported under EMIR has significantly improved following the EMIR REFIT reform in 2024. ESMA has moved from testing generative artificial intelligence to operational use in internal activities, supervisory analysis, and market abuse projects. The authority is working to reduce unnecessary reporting by eliminating duplications, the "report once" principle, and a more integrated framework for fund reporting.
The ESMA report covers data reported under several European regimes: EMIR for derivatives, SFTR for securities financing transactions, MiFIR for transactions and market transparency, AIFMD and MMFR for funds, DORA for major digital incidents, ESMA registers, prospectus reporting, securitizations, missing sales, central counterparties, rating agencies, and crowdfunding.
For investors, this data is not an administrative matter. It is used for detecting market abuse, monitoring investment firms, assessing risks from funds, analyzing links between banks and the non-bank financial sector, verifying derivatives markets, and evaluating digital incidents that may affect financial infrastructure.
ESMA states that, in 2025, it pursued two parallel objectives: more efficient use of supervisory data and reducing unnecessary complexity for reporting firms. The authority launched a consultation on June 23, 2025, regarding the simplification of financial transaction reporting, focusing on the EMIR, MiFIR, and SFTR regimes.
Two directions are being analyzed: eliminating existing duplications and a more structural change based on the "report once" principle, whereby firms would submit information only once, and authorities would share data among themselves. In this context, ESMA decided not to propose changes to the main MiFIR reporting frameworks until the completion of a broader analysis, to avoid further uncertainty for the market.
In the area of funds, ESMA is working on an integrated reporting framework to reduce fragmentation between AIFMD, UCITS, MMFR, ECB reports, and national reports. The goal is to harmonize definitions, timelines, and required data, so that authorities receive comparable information, and the industry avoids parallel reporting.
ESMA also organized its first Data Day on December 5, 2025, with authorities and representatives from the financial industry. Discussions focused on simplification without deregulation, reporting data once, data sharing between authorities, and the quality of data necessary for supervision and for the use of artificial intelligence.
Artificial intelligence has transitioned in 2025 from experiments to operational use within ESMA. The authority has expanded the use of tools such as M365 Copilot, improved security measures for sensitive content, and worked on AI applications tailored to supervisory activities.
One project uses RADAR data about rating agencies to verify elements related to their methodologies. The system searches for information in communications, compares it with the metadata reported by rating agencies, and signals possible methodological inconsistencies.
ESMA also conducted a pilot project, together with national authorities, for the use of AI in detecting behaviors that may indicate market abuse. The project mapped existing research, designed AI models for certain patterns of abuse, and tested them in a controlled environment. ESMA is now analyzing the results together with national authorities.
The report also shows how large the data infrastructure of financial markets has become. In 2025, the volume of transactions reported under MiFIR reached approximately 9.2 billion, up from 7.7 billion in 2024. Approximately 4.7 billion of these transactions were reported through approved reporting mechanisms.
MiFIR data is used for market supervision, detecting market abuse, monitoring retail investor activity, analyzing market structure, and calculating transparency indicators. ESMA shows that transaction data is now significant enough to replace other reporting streams in certain transparency calculations, which reduces the burden on trading venues.
The quality of EMIR data, used for monitoring the derivatives market, has significantly improved following the transition to EMIR REFIT in 2024. Indicators for transactions and unreconciled positions have decreased from levels of approximately 30% to 55% in 2024 to about 10-12% by the end of 2025.
The report indicates improvements in data regarding valuations and contract maturity. Outdated valuations have decreased from approximately 25% in mid-2024 to below 5% by the end of 2025. Fields with uncompleted or abnormal maturity data have decreased from about 11% to around 2%.
For EMIR, authorities use data for monitoring derivatives, assessing systemic risk, verifying clearing obligations, analyzing energy markets, CDS exposures, exposures to non-EU counterparties, and links between banks and the non-bank financial sector. In crises, authorities can use this data to quickly see where exposures and contagion risks are.
For SFTR, data on securities financing transactions shows improvements but also persistent issues. Indicators for incorrectly reported transactions and unmatched reports decreased in 2024-2025, but matching loans and collateral remains challenging. The report shows that, by the end of 2025, collateral matching was still at high error levels, around 65%.
SFTR data helps authorities see how repo and securities lending markets function, where collateral is located, how banks are connected to non-bank financial institutions, and what vulnerabilities may arise in secured financing. In 2024, no administrative sanctions were imposed in the EU for SFTR violations.
In the area of alternative funds, the report shows that 28,894 alternative investment funds reported to ESMA for the second half of 2025, with a total net asset value of 7.5 trillion euros. The data is used for risk monitoring, leverage analysis, stress testing, and assessing UCITS funds' indirect exposures to riskier assets.
For money market funds, the database includes 375 funds, with a total net asset value of 1.9 trillion euros. MMFR reports are used for monitoring liquidity, portfolio composition, stress test results, and the industry's resilience to market shocks.
DORA introduces a new reporting area for major digital incidents. Financial institutions must notify major ICT incidents, send interim updates, and submit a final report after resolving the cause. ESMA is already observing differences in interpretation between entities and is preparing clarifications for more uniform application of the criteria.
In the first year of DORA reporting, ESMA focused on data availability and their transmission under confidentiality conditions. In 2026, European authorities will publish the first joint annual report on major ICT incidents, and ESMA is working on an automated validation and dissemination tool for these reports.
The report also covers the European Single Electronic Format (ESEF), through which listed issuers publish annual financial reports in a digital format. ESMA has collected 2,745 annual financial reports for the financial year 2024, but notes that the set should not be interpreted as exhaustive, as the collection process continues.
For Romania, the ESEF table indicates 44 annual financial reports collected for 2024, compared to 30 for 2021 and 35 for 2022, while for 2023 there is no number in the current set collected by ESMA. The report explains that coverage also depends on the capacity for automatic and manual collection of documents.
In securitizations, the remaining principal amount due for asset-backed securities reached 885 billion euros by the end of 2025, compared to 826 billion euros a year earlier. 59% of these amounts are related to residential mortgage loans, followed by corporate loans, auto loans, and consumer loans.
For missing sales, ESMA shows that all national authorities have switched to daily reporting since the end of 2024. The new automated verification framework has significantly reduced missing reports: from 2,089 authority-day combinations with missing data at the beginning of 2025 to 123 at the beginning of 2026. For Romania, the table shows 58 missing reports on January 3, 2025, and zero on February 10, 2025, and 19 on January 19, 2026.
ESMA also created a network of experts in supervisory technology, SupTech, in 2025, together with national authorities. The network has identified over 400 SupTech projects and has over 300 national experts on the distribution list. More than 20 projects or tools in source code form have been made available to network members.
The report shows that the future of European financial supervision will depend on three directions: cleaner data, less repetitive reporting, and more intensive use of technology. ESMA aims to expand the publication of open-source code on GitHub, use more generative AI, and automate quality controls so that issues can be detected more quickly.
ESMA is the European authority responsible for the supervision and coordination of securities markets in the European Union. The data it receives from firms, exchanges, registers, funds, rating agencies, and national authorities are used for investor protection, financial stability, market integrity, and the orderly functioning of transactions.
The 2025 report shows that financial supervision no longer relies solely on rules but also on the authorities' ability to transform large volumes of reporting into useful signals. The quality of data determines whether an anomaly is detected in time, whether a market is properly supervised, and whether firms that report incorrectly are treated the same across all member states.
https://2eu.brussels/ro/stiri/esma-foloseste-inteligenta-artificiala-si-date-mai-curate-pentru-a-supraveghea-pietele-financiare-din-ue
Latest News
10:22
10:08
10:04
09:56
09:32
See more news