Plaid pulled in $546 million in annual recurring revenue in 2025, which is a 40% increase from $390 million in 2024. Plaid’s growth has expanded beyond linking bank accounts. New product lines, which include identity verification, fraud detection, payments initiation, and credit underwriting, now account for more than 20% of total revenue. Plaid Protect grew 400% year-over-year in 2025. Payments facilitation grew 250%. These new products grow at rates of 90% or higher, which pulls the blended company growth above the 40% mark. Core account-linking revenue grows slower than 40% because banks and card networks build competing open-banking rails. You know the core connectivity layer faces intense pricing pressure. Plaid’s revenue architecture relies on usage-based pricing across two main models. The platform uses one-time fees and subscription fees. It also applies per-request fees for every successful API call. Enterprise customers often move to custom tiers to negotiate volume discounts. These tiered structures allow Plaid to capture more value from high-volume fintech apps. This $546 million ARR comes from more than 150 million connected consumer bank accounts across 7,000 fintech apps like Robinhood and Venmo. Plaid’s growth is fueled by the expansion into identity and payments. The company’s growth reaccelerated after a period where growth slowed to 12% in 2023. This pivot to a multi-product platform aims to move Plaid from a data aggregator to a core infrastructure provider.
Competitive pressure from banks and card networks
Large banks like JPMorgan Chase changed the unit economics for aggregators in 2025 by charging fees for API access. JPMorgan Chase introduced a $1.50 fee per authorization token for each data request. This pricing model compresses the gross margins of pure-play aggregation platforms. Plaid is an intelligence platform. It must cross-sell risk and analytics products to offset these direct bank API access fees. For developers, the platform provides a free Sandbox environment with up to 200 API calls. Commercial users pay for Auth products between $0.10 and $0.25 per successful call. Accessing historical transaction data through the Transactions product ranges from $0.30 to $0.60 per call. Identity verification costs scale between $0.15 and $0.30 per call. MX Technologies remains a smaller competitor with roughly $95 million in annual recurring revenue. Envestnet Yodlee remains a legacy player with low growth. Finicity is owned by Mastercard and Tink is owned by Visa. Akoya belongs to an 11-bank consortium. The competitive field is crowded with players specialized in different niches. Finicity focuses on mortgage and income verification with direct agreements at major banks like Bank of America and Wells Fargo. MX Technologies specializes in transaction enrichment and categorization for lending and wealth products. Plaid remains the leader for developer experience and coverage of the long tail of smaller institutions.
| Product Category | 2025 Growth Rate |
|---|---|
| Plaid Protect | 400% YoY |
| Payments Facilitation | 250% YoY |
| New Products Total | >90% YoY |
| Blended Total ARR | 40% YoY |
Valuation and the AI agent expansion
The $8 billion valuation reached in the February 2026 tender offer remains much lower than the $13.4 billion peak recorded in 2021 because the massive contraction in fintech multiples followed the era of high interest rates. Plaid uses its connections to power AI agents for major tech firms. It launched a partnership with OpenAI in May 2026 and integrated with Meta’s Muse AI agent in September 2026. These connections allow users to share balances and transactions with AI models, but the Meta connection is read-only and requires explicit user approval. These AI tools help users ask about spending and investment patterns. The $8 billion valuation implies a 15x ARR multiple, which is a recovery from the $6.1 billion valuation in April 2025. In 2025, Plaid reached adjusted EBITDA profitability for the first time, which marks a departure from previous years of cash burn. The company’s scale is massive, with 52% of US adults using at least one open banking service in 2025. This user base provides the data needed to train proprietary machine-learning models. In May 2026, Plaid deployed dual foundational AI models for transaction classification and sequential analysis. These models boosted categorization accuracy by 13%. Will the AI-driven revenue offset the shrinking margins of the core data layer?
