The US is the largest economy without integrated Digital Public Infrastructure

Digital Public Infrastructure has become the canonical term for the framework that combines identity, real-time payments, and consent-based data exchange into a single interoperable stack. The G20 formally adopted it under India’s 2023 presidency. The World Bank and UNDP made it the basis of joint guidance, publishing the DPI Playbook and the DPI SDG Compendium (UNDP G20 DPI press release; World Bank DPI publication). The BIS Innovation Hub built two of its largest live projects (Nexus on payments interlinking, Mandala on cross-border compliance) on the DPI premise that the three layers gain economic value from coordination. By mid-2026 the international policy conversation has settled on integration, and the US is conducting its own conversation separately.

The settled position is not a soft preference. India has signed agreements with 23 countries (as of February 2026) to share its DPI stack, from Sri Lanka to Jamaica, and the Union Budget 2025-26 funded deeper exports through the India-AI mission (A Junior VC on India Stack export). The European Union’s Digital Identity Wallet enters mandatory roll-out across all member states on 24 December 2026 under eIDAS 2.0, with regulated entities required to accept by end of 2027 (European Commission EUDI Wallet page). Singapore’s Singpass connects 4.5 million users (97% of the eligible population) to more than 2,000 public and private sector services, and PayNow is in active use by 80% of consumers and businesses, accounting for 99% of Singapore bank transfers (World Bank Singapore blog; PwC Singapore Payments State of Play 2026). Australia’s Consumer Data Right, New Payments Platform, and Trust Exchange operate under a single Treasury-coordinated framework with $650M committed in the 2026 federal budget to Digital ID expansion (CDR rollout; Biometric Update on Digital ID budget; Ashurst on Trust Exchange). The US has FedNow, RTP, §1033, mDL, and Login.gov, all running on separate regulatory tracks under different agencies, with no coordinating body and no interoperability mandate.

The economic case for integration sits in the complementarities between the three layers. Identity reduces the fraud cost of real-time payments because the rail can settle in milliseconds against a verified counterparty rather than against an unauthenticated pseudonym. Open banking gives real-time payments a use case beyond person-to-person transfer because every consent-permissioned data action becomes a potential settlement instruction. Real-time payments give open banking a settlement leg because consent-permissioned data access becomes useful only when paired with instant value transfer. Built separately, each layer carries its own fraud cost, its own onboarding friction, and its own consumer-adoption curve. Built together under a single regulatory framework, the three exhibit increasing returns. The integrated stack reduces unit cost; the fragmented stack pays the unit cost three times over.

India operates the largest live deployment. UPI processed 21.7 billion transactions in January 2026 alone, worth roughly $340 billion at month-end exchange rates (India Policy Hub on India Stack). Aadhaar covers 1.44 billion residents. DigiLocker has 676 million users storing 9.5 billion digital documents. The Account Aggregator framework runs the consent layer for financial data, with the Reserve Bank of India and SEBI as the coordinating regulators. The numbers are large by any measure and disproportionate by population: UPI’s monthly volume now exceeds the combined annual volume of FedNow and RTP for any month in 2025. Critics of the Indian model focus on Aadhaar’s mandatory enrolment for welfare access and on the Indian Supreme Court’s earlier rulings on its constitutional limits. These critiques are fair as applied to India’s institutional choices, and the US would not adopt them. The Indian implementation is one of several available models, and the alternative versions in Singapore and Australia are closer to the institutional tradition the US would actually adopt.

Singapore demonstrates the alternative. Singpass is opt-in, federated, and governed under the Personal Data Protection Act. The Monetary Authority of Singapore and the Government Technology Agency built SGFinDex on top of Singpass for consumer financial data exchange, with the Association of Banks in Singapore, the Life Insurance Association, and 15 financial institutions as participants (MAS SGFinDex page). Roughly 290,000 bank accounts have been connected through SGFinDex, with 620,000 data retrievals completed. PayNow’s 80% adoption sits on the same identity layer. The model is closer to the US legal tradition than to India’s: opt-in, consent-explicit, with the regulator coordinating across the three layers but issuing rather than mandating. Singapore is a small market by population and a poor comparator for absolute volume, but the institutional model travels.

Australia is the closer comparator on regulatory tradition and economic size. The Consumer Data Right, established under the Treasury and the Australian Competition and Consumer Commission, runs consent-based data sharing across banking, energy, and shortly non-bank lending; Version 8 Rules extend the regime to non-bank lenders from July 2026. The New Payments Platform processed nearly 2 billion real-time payments in 2025 with $7 billion in daily volume across 128 million accounts able to send and receive, and PayID registrations have passed 27 million (Australian Payments Plus NPP page). The Australian Digital ID Act 2024 established a national Digital ID System, and the Trust Exchange opens to private-sector accreditation on 30 November 2026. The 2026 federal budget committed $650 million to Digital ID expansion. Each layer is being built by a different operational body, but the regulatory coordination sits with the Treasury and the RBA’s Payments System Board, and the policy framing is integration.

The US has the same three layers in various states of progress, but no coordination. FedNow attracted 1,600 financial institutions by early 2026, 500 of them added in 2025, and moved $853.4B in 2025 transaction value across an average of 30,000 daily transactions (Federal Reserve FedNow volume and value statistics). The Clearing House’s RTP network reached 1,135 institutions and $1.3 trillion in 2025 volume (Financial Brand on FedNow and RTP). Combined, US instant-payments rails moved roughly $2 trillion through about 2,700 institutions. The institutional adoption is real. The consumer-facing experience is not on the rails the Fed and TCH built; it is on Zelle, Cash App, and Venmo, which sit on top of the rails as bank-owned or non-bank-owned consumer products. The integrated identity layer that Singpass or Aadhaar provides at the underlying rail level does not exist in the US, and the consumer-protection regime that backs the rails in Australia and the UK has no US equivalent for FedNow.

§1033 was the closest the US came to coordinating the data layer. The CFPB issued the final rule in October 2024, with a phased compliance timeline starting at Tier 1 banks in April 2026 and reaching 400 million consumer accounts (Federal Register §1033 final rule). The rule was stayed and reopened for reconsideration under the new CFPB leadership in 2025. As of May 2026 the April 2026 compliance date is effectively dead. The Financial Data Exchange has 114 million consumer accounts on its APIs and remains the de facto US open-banking standard (Banking Dive on CFPB picking FDX), but FDX is an industry-coordinated body rather than a regulator-issued framework, and its identity-credential product does not yet exist.

State-issued mobile driver’s licences are the most mature US example of one DPI layer shipping in the absence of a federal framework. Twenty-one states plus Puerto Rico run live mDL programmes (AAMVA mobile driver license page); ISO 18013-7 has standardised online presentation; the schemes have started integrating around state mDL rather than building parallel rails. The mDL story is its own argument, treated separately. For DPI purposes, the relevant fact is that state mDL has shipped without coordination with FedNow, RTP, §1033, or Login.gov, and there is no roadmap to connect them.

The fragmentation premium is a cost the US private sector pays for that institutional choice. Banks carry duplicate identity-verification stacks, each tied to separate KYC vendors and document-image processors, while the fintech layer above them carries duplicate data-access integrations with Plaid, MX, Akoya, and FDX-direct connections multiplying across institutions. Merchants pay for separate CNP risk-scoring, behavioural biometric, device-fingerprinting, and dispute-defence stacks, and consumers pay the resulting costs through higher acceptance and credit prices that bake in the duplication. The Atlanta Fed Payments Forum’s October 2025 publication noted that stringent identity verification operates as a barrier to access for the unbanked and underbanked (Atlanta Fed Payments Forum, October 2025); the fragmentation premium is regressive in its incidence as well as inefficient in its level.

The pro-fragmentation argument holds that this cost is the price of innovation, and that integrated stacks in Asia produce less product-level competition than the US fragmented stack does. The argument depends on what is measured. US innovation in customer-facing fintech products is real. US innovation in core payments infrastructure (instant payments, identity, open banking) lags Asia and Europe by every available metric, including consumer adoption, fraud rates, and cross-border payments share. The two are not mutually exclusive. An integrated DPI substrate at the infrastructure layer does not preclude private-sector competition at the product layer. The Singapore model demonstrates this, with Singpass as substrate and private-sector competitors layered on top. The pro-fragmentation argument confuses substrate fragmentation with product competition.

Another objection treats the US as the global payments leader by absolute volume, which makes the comparative framing misleading. The US handles the largest absolute volume of card and CNP payments globally, and is the home market for Visa, Mastercard, PayPal, and Stripe. The objection is true and irrelevant. Absolute volume measures market size, not infrastructure quality. The relevant comparison is in how each piece is built: integrated stack versus fragmented, public-utility versus private-sector, interoperable versus siloed. On those axes the US trails, and the gap is widening as Australia, the EU, and the Indian export programme institutionalise the integrated approach.

The most serious objection comes from civil-liberties commentators, including the EFF, the Electronic Privacy Information Center, and Bruce Schneier, who have argued that integrated digital identity enables state surveillance and that India’s Aadhaar architecture (mandatory enrolment for welfare access, biometric primary key, centralised authentication) is the natural endpoint. The argument is correct about Aadhaar and wrong about DPI. The Singapore implementation is opt-in, federated, and governed under PDPA. The Australian regime is consent-by-default, with privacy-by-design principles written into the CDR statutory framework. The EUDI Wallet under eIDAS 2.0 is built around the user holding their own credentials in a wallet on their device, with selective disclosure. The relevant comparators for the US legal tradition are Singapore, Australia, and the EU, not India. The US debate has conflated DPI as a category with Aadhaar as one implementation, and the article corrects that conflation.

The political-economy implication is that the US private sector should be lobbying for integration, not against it. The fragmentation premium is a cost the private sector pays, not a benefit it captures. The actors that carry it are diffuse: banks running duplicate identity stacks, the fintech layer carrying per-institution integration overhead, merchants maintaining separate fraud stacks. Beneficiaries of US fragmentation are narrow: data brokers whose business model rests on SSN-as-identity, KYC vendors whose revenue depends on per-onboarding fees, and a handful of legacy verification incumbents. The political coalition for integration is large; the coalition against it is small. The US has not built the integration anyway, because the coordination problem across federal agencies has been politically unsolved.

The G20 DPI consensus is the international policy convergence the US is operating outside of. When the framework moved from theoretical to operational under joint endorsement by the G20, World Bank, and UNDP in 2023, the international conversation settled. India is exporting it. Brazil’s Drex and Pix architecture is building a parallel version. The EU is mandating it. The UK’s One Login is the federal-identity piece. Singapore and Australia run the institutionally mature versions. The US has not endorsed the framework as a domestic strategy and has no roadmap to do so. The cost of operating outside the consensus is measurable now in fraud-cost differentials and consumer-adoption rates, and it will become more measurable as cross-border payments share migrates to integrated rails over the next decade.

For private-equity diligence on US payments and fintech assets through 2030, the binding question is not whether the asset has a DPI strategy. The question is whether the asset’s economics survive a US institutional move toward DPI at any point in the next ten years. Assets whose business models rest on the fragmentation premium (identity-verification vendors selling per-onboarding fees, data brokers selling SSN-keyed identity products, KYC consultancies selling per-bank integration) face revenue compression in any DPI scenario. Assets that integrate with the underlying rails (FedNow and RTP processors, FDX-connected fintechs, mDL-aware verification vendors) have optionality. The diligence question for the next decade is whether the asset is short or long the fragmentation premium, and the public-market analyst community is not yet pricing that exposure.