Chainlink, the widely used oracle platform, announced today the results of the second phase of an industry initiative aimed at standardizing and speeding up corporate actions processing by combining blockchain, Chainlink’s oracle technology, and artificial intelligence.
With 24 leading financial market participants taking part, the effort produced a production-grade proof that structured, validated corporate actions data can be delivered into legacy systems in minutes rather than days, a change that could save the global financial system tens of billions of dollars annually and sharply reduce settlement errors and operational risk.
Corporate actions, everything from dividends and rights issues to more complex events, currently cost the industry an estimated $58 billion a year, a figure rising roughly 10% year over year while automation rates have slipped below 40%. Citi’s 2025 Asset Servicing report, cited by the initiative, notes that the average corporate-action event now touches more than 110,000 firm interactions and costs about $34 million to process, with 75% of market participants still relying on manual revalidation of data.
From LLM Experiments to Production Deployment
Phase 1 of the project had already demonstrated that large language models, including OpenAI’s GPT series, Google’s Gemini, and Anthropic’s Claude, could extract structured fields from unstructured corporate-action announcements and publish unified “golden records” onchain. Phase 2 built on that work, delivering a production-grade deployment that improved speed, reach and accessibility across both traditional infrastructure and blockchain environments.
Chainlink’s Runtime Environment (CRE) was used to orchestrate validation across multiple AI model outputs. Confirmed results were converted into ISO 20022-compliant messages and transmitted over the Swift Network. Simultaneously, Chainlink’s Cross-Chain Interoperability Protocol (CCIP) distributed the same verified records across the DTCC’s blockchain ecosystem and other public and private ledger environments, ensuring the same attested record is available to custodians, post-trade systems and smart contracts.
The initiative also introduced institutionally designated roles for data attestors and contributors who cryptographically attest to data accuracy and fill in missing fields, creating a verifiable chain of custody across a corporate action’s lifecycle. During testing, the system reached nearly 100% consensus among the AI models on all evaluated corporate actions and demonstrated reliable multilingual support (including disclosures in Spanish and Chinese), widening the solution’s global applicability.
Why it matters
Participants span market infrastructures, custodians, banks and asset managers. Named organisations include Swift, DTCC, Euroclear, SIX, TMX, CEVALDOM, Grupo BMV, ADDX, Orbix Technology, Marketnode and Wamid, along with UBS, DBS Bank, BNP Paribas Securities Services, ANZ, Wellington Management, Schroders, Zürcher Kantonalbank, Vontobel, CTBC Bank, Causeway Capital Management, Sygnum Bank, AMINA Bank and Zand Bank.
By delivering an attested, real-time “golden record” for corporate actions that is accessible onchain and across legacy networks, the initiative aims to remove longstanding friction: late or inconsistent data, manual reconciliation, and the resulting settlement fails. For tokenized equities, a growing class of digital assets, the unified record provides a common reference that smart contracts and custodians can rely on, enabling higher levels of automation and better synchronization across onchain markets.
Sergey Nazarov, Chainlink’s co-founder, framed the result as a milestone: solving the corporate actions data validation problem “using AI Oracle Networks” and placing consensus-validated records onchain “is a huge step forward for how corporate actions can work across the capital markets.”
Leaders from participant organizations emphasized interoperability, trust and industry coordination. Dan Doney, Managing Director & CTO of DTCC Digital Assets, highlighted how distributed ledger technology can boost transparency and connectivity. Euroclear’s Stéphanie Lheureux said scalable digital market infrastructure must align with systems institutions already trust, a point she tied to the cross-industry cooperation demonstrated in the initiative.
Asset-servicing and banking executives also stressed client impact: Wellington Management’s Mark Garabedian called faster, standardized delivery of accurate corporate actions data “essential” for reducing operational overhead and reacting to time-sensitive events.
BNP Paribas’ Wayne Hughes said the project offers a chance to see how blockchain can improve corporate actions processing for clients. Several participants underlined broader market benefits: improved efficiency and lower costs, stronger auditability and governance, and a foundation to expand tokenized asset servicing at scale.
Next Steps
The collaborators say future work will focus on supporting more complex corporate actions (for example, stock splits), expanding jurisdictional and currency coverage, and strengthening privacy and governance controls to match operational and compliance requirements across global institutions.
The group has published a full report titled Establishing a Unified Standard for Asset Servicing With the Chainlink Platform, Blockchains, and AI, which outlines the technical architecture, test results and proposed next steps for industry adoption. Fragmented corporate actions processing has been a chronic drag on post-trade efficiency.
This Chainlink-led initiative, backed by major market infrastructures, custodians, banks and asset managers, shows a practical path to convert unstructured announcements into validated, standardized records that can be consumed simultaneously by both legacy systems and blockchain platforms. If broadly adopted, the approach could markedly lower costs, shrink settlement times, and clear a major obstacle to scaling tokenized public equities.
from Crypto News: Bitcoin, Ethereum & Price Data | BlockchainReporter https://ift.tt/KImOAbu
via IFTTT
