UPI's Hidden Plumbing: Why Bank Concentration Matters
12 Jul 2026
India's UPI network just crossed another milestone — over 2,272 crore payments in June 2026 — but beneath the headline growth number sits a plumbing story that matters more to founders building on top of the rails: a small number of apps and an even smaller number of sponsor banks are doing most of the work.
The concentration problem
PhonePe and Google Pay together handle roughly 80% of all UPI volume. That's a striking level of concentration for a public payments utility, and it means most of India's digital transaction flow is routed through just two consumer-facing apps.
Behind those apps sits an even tighter layer: sponsor banks. PhonePe's payment handles run on Yes Bank, Axis, and ICICI. Google Pay's run on Axis, HDFC, ICICI, and State Bank. Yes Bank in particular stands out — its share of incoming UPI payments has roughly doubled over the past two years, and the report describes it as the dominant sponsor bank for the largest merchant apps.
That combination — two apps carrying most of the traffic, and one bank increasingly central to settlement — is the kind of structural pattern that draws attention in infrastructure-heavy networks. If Yes Bank or either of the two lead apps faced disruption, the report notes, it could represent a single point of failure risk for a system now processing billions of transactions a month.
A challenger's fast climb
Not everything in the numbers points to entrenchment, though. Super.money, a Flipkart-launched app from 2024, moved from outside the top fifty UPI apps to the top five in about a year. That kind of movement suggests the market, while concentrated, isn't fully closed to new entrants with enough distribution muscle behind them.
The P2P-to-P2M shift
Another data point worth tracking: person-to-merchant (P2M) payment volumes overtook person-to-person (P2P) volumes back in 2022. That crossover has held since, and it points to UPI increasingly functioning as commerce infrastructure rather than just a peer-transfer tool — a shift that matters for anyone building merchant-facing payment products.
The paisa-level detail
One small but telling mechanic: when both the payer and payee use the same sponsor bank for a transaction, that bank saves about one paisa in resolution fees. It's a tiny number per transaction, but at UPI's scale — billions of transactions monthly — it's the kind of micro-economics that can shape how sponsor banks and payment aggregators think about routing and infrastructure design.
Why founders should care
- If your product distribution leans entirely on PhonePe or Google Pay integrations, there's a reasonable chance that a disruption or policy change affecting either app could meaningfully impact your transaction flow, given they likely carry the bulk of consumer UPI traffic.
- Sponsor bank concentration — particularly Yes Bank's growing role — suggests founders building payment infrastructure may want to evaluate multi-bank redundancy rather than depend on a single settlement partner.
- Super.money's rise indicates that UPI market share is likely still contestable; well-capitalized new entrants may plausibly gain meaningful share within a year, not a decade.
- The 2022 P2P-to-P2M crossover likely signals a maturing merchant payments market, which may be worth monitoring for commerce and checkout-focused startups.
- The paisa-level fee savings from same-bank transactions could point to a narrow but real optimization opportunity for high-volume payment aggregators willing to build routing logic around sponsor-bank alignment.
What's still unclear
The report leaves several open questions: which sponsor banks power UPI apps beyond PhonePe, Google Pay, and Super.money; what specifically drove Yes Bank's share to double; how the paisa-level fee savings mechanism affects UPI economics at scale; whether regulators have flagged sponsor-bank concentration as a concern; and what exactly enabled Super.money's rapid climb. Founders operating in this space should watch for clarity on these points as the ecosystem evolves.