India’s payment establishment has picked a side in the agent-payments race: the public one. Rather than letting Visa, Mastercard, Stripe, and a queue of fintechs each roll out proprietary schemes for AI agents buying and paying on their own rails, Indian authorities want agent-to-agent transactions on UPI standardized as a public good. The distinction matters. Most agent-payment infrastructure being built today is corporate by construction — closed loops where the network operator also writes the authorization rules. India is arguing that a payment layer for autonomous software should look more like a utility.
What India Is Actually Proposing
The reported position is that AI agents paying on UPI should operate as a public standard, not as a set of competing private protocols. Concretely, this means the interoperability layer — how an agent identifies itself, proves authorization, and initiates a payment — would be defined openly on top of UPI rather than owned by whichever card network or processor gets there first. India has run this play before. UPI itself was built as public infrastructure that forced private wallets into interoperability, and it demolished the closed-loop wallet business in the process. Applying the same logic to agent payments is less a novel ideology than an extension of an existing regulatory philosophy.
The Corporate Scramble It Pushes Against
Elsewhere, agent payments are consolidating around corporate alliances. Card networks have joined coalitions drafting agent-payment standards; banks like DBS are pairing with Stripe to build bank-grade rails for shopping agents; and protocol-level efforts like x402 are betting on stablecoin settlement over open HTTP extensions. These approaches share a premise: whoever controls the identity and authorization layer for spending agents controls the volume. India’s framing rejects that premise outright. A public standard on a state rail removes the toll booth before it is built. The open question is whether global agent commerce — which is largely dollar-denominated — can be routed through a rupee-settling public rail at all.
Where Stablecoins Fit
Here the story gets interesting for this audience. UPI settles in rupees through supervised banking channels. Agent commerce, as observed on chains like Base and the XRP Ledger, defaults to dollar-pegged stablecoins because software wants a unit of account that is stable, programmable, and globally legible. A public UPI standard for agents and stablecoin-denominated agent rails are not necessarily competitors — they may be parallel stacks, one domestic and one cross-border. But if India’s public-standard model proves that open agent-payment rules can scale without a corporate owner, it strengthens the argument that standards like x402 should stay genuinely open rather than drift toward consortium capture.
What to Watch
The first concrete test will be technical specifications: does NPCI or the RBI publish an actual agent-identity and authorization standard, or does this remain a policy posture? Second, watch whether multinationals operating in India are required to route agent transactions through the public layer — that would be the UPI playbook replayed. Third, observe whether dollar-settling agent volumes around India simply bypass UPI via stablecoins on Tron or Ethereum, as remittance flows already do. Intent is easy; enforcement against software that can settle anywhere is the hard part.