Chainlink and a coalition of 24 financial institutions, including Swift, DTCC, Euroclear, UBS, DBS Bank, and BNP Paribas Securities Services, have completed Phase 2 of an initiative using AI models and blockchain to process corporate actions data. The group reached near-100% consensus on the data, targeting an industry problem that costs $58 billion a year.
Chainlink announced the Phase 2 results on September 29, 2025, revealing that the 24-institution coalition achieved near-100% consensus on corporate actions data processed through a combination of AI models and blockchain infrastructure. Corporate actions cover dividends, mergers, stock splits, and other events companies impose on shareholders.
A $58 billion back-office problem
Corporate actions cost the global financial industry an estimated $58 billion a year, a figure climbing 10% annually. A single corporate event can generate costs as high as $34 million and involve over 110,000 interactions among firms. The reason: less than 40% of corporate actions processing is automated, leaving staff to manually pull data from PDFs, foreign-language press releases, and regulatory filings before typing it into systems.
AI models check each other's work
Phase 2 combined large language models from OpenAI and Google with Chainlink's oracle infrastructure into a multi-step verification pipeline. Instead of trusting one model's output, the system runs corporate actions data through multiple models and checks their answers against each other, with the results also confirmed through institutional attestation rather than the models grading their own work.
The processed data comes out in structured records compliant with ISO 20022, the messaging standard finance has been migrating toward. Distribution runs two ways: through the Swift network for traditional infrastructure, and through Chainlink's Cross-Chain Interoperability Protocol for blockchain-native systems. The pipeline also converts source documents written in multiple languages into standardized records within minutes.
Why the participant list matters
Swift processes messaging for over 11,000 financial institutions globally, DTCC settles the vast majority of US securities transactions, and Euroclear is one of the world's largest settlement systems. None of these are crypto-native startups experimenting with the technology — they are institutions that run global financial infrastructure, and they are building with Chainlink's tools.
The data plumbing also serves tokenization: smart contracts managing tokenized stocks need accurate, structured corporate actions data — notice of an upcoming split, for instance — to function correctly. Chainlink says the initiative builds a single source of truth accessible to both smart contracts and traditional systems.
If even a fraction of the $58 billion in annual costs shrinks through automation, the savings flow directly to the institutions adopting the technology, positioning Chainlink's infrastructure as middleware between AI, blockchain, and traditional finance.
Source: Crypto Briefing
Trading involves risk.
Risk Disclaimer
