Plaid Didn't Build Most Connectors
Tony Xiao, founder and CEO of Venice, on the opportunities in financial data aggregation
The connector layer in bank aggregation was far less proprietary than it looked, which meant Plaid won more on packaging than on unique raw coverage. The hard part was keeping thousands of fragile bank connections alive, especially for small institutions that generated little revenue. Evidence across related research shows the leading aggregators built the top tier themselves, then filled out the long tail through contractors, partners, or licensed connectivity, which is why coverage between Plaid, Yodlee, Finicity, and MX overlapped so heavily.
-
What fintechs actually bought from Plaid was often not more banks, but a cleaner developer workflow. Plaid paired pay as you go pricing and simple docs with account linking, balance checks, and transaction history, while Yodlee was earlier but harder to integrate and sold more like enterprise software.
-
This also explains why adding a second aggregator often did not double usable coverage. Prior work found the biggest players built proprietary links for only the top roughly 500 to 1,000 institutions, while the long tail of 11,000 plus banks was frequently outsourced or resold, leaving large overlap across providers.
-
The strategic value shifted upward from raw connectors to products built on top of the data stream. As the connector layer commoditized, Plaid used its distribution across thousands of fintechs and hundreds of millions of linked accounts to sell adjacent products like identity, risk, ACH, and enriched transaction data.
The next phase of aggregation belongs to companies that either secure direct API partnerships or turn commodity connectivity into higher value workflows. Plaid has already been moving in that direction through open finance partnerships and network products, and newer infrastructure companies are following the same playbook, using connectors as the wedge and monetizing the software layer above them.