Picture a US bank trying to prove to a regulator that it has flagged every transaction above the $10,000 reporting threshold, without showing the regulator the underlying transactions. That trade-off is what zero-knowledge proofs are built to make possible. According to a 2023 Bank for International Settlements working paper on privacy-enhancing technologies in CBDCs, ZK schemes are now the most discussed cryptographic primitive in central bank and tokenisation pilots. The questions are how, when, and at what cost they reach US finance.
How zero-knowledge proofs entered the US financial conversation
The maths is older than most of the people working on it. Goldwasser, Micali and Rackoff defined zero-knowledge proofs in 1985 as a way to prove a statement is true without revealing why. For three decades the work stayed academic. Zcash made the first production deployment in 2016 when it shipped zk-SNARKs to a public blockchain. By 2020, Ethereum’s scaling roadmap had pulled ZK rollups into the mainstream conversation, and Wall Street started paying attention shortly after that.
The US entry point has not been crypto-native, though. It has been compliance. A bank that wants to share a fraud signal with another bank without leaking customer identities, or prove to OFAC that none of its counterparties are sanctioned without exposing its book, has a problem that ZK can solve cleanly. The Federal Reserve Bank of Boston flagged the privacy-versus-supervision tension in its Project Hamilton work on a digital dollar prototype, and the same tension is now showing up in commercial pilots.
That route in matters. It means the early US adopters of ZK are not protocol teams chasing throughput. They are compliance officers and risk leaders trying to solve a regulatory problem with a cryptographic answer. The conversation has shifted from how fast a proof can be generated to whether a proof can be used as evidence.
What the adoption signal looks like in 2025
Adoption numbers in ZK are still small relative to the broader fintech market, but the curve is steepening. According to L2Beat’s scaling summary, as of 2025 more than 11 active ZK rollups carried over $1 billion in combined total value secured, up from under $1 billion two years earlier. The growth is concentrated in a handful of rollups, but the institutional flows behind them now include tokenised treasuries and money market funds, not just retail trading.
The pattern matters because it suggests the cryptography is no longer the bottleneck. Prover costs have fallen by roughly an order of magnitude over the past three years as hardware acceleration and recursive proof composition have matured. The remaining drag is regulatory: most US banks will not run a privacy-preserving rail without a clear signal from their supervisor that the audit trail still works. That signal is starting to arrive.
Vendor activity tells the same story from a different angle. Filings on the OCC docket, conference panels at SIFMA and Davos Blockchain events, and the volume of ZK-related procurement RFPs at US Tier 1 banks have all stepped up. None of this is the kind of growth that generates a headline. It is the slow, steady accumulation of pilots that historically precedes a real shift in infrastructure spending.
Where US regulators have started to engage
The Treasury Department’s Office of Financial Research has published research notes on privacy-preserving analytics and on the risks of opaque on-chain systems. The Securities and Exchange Commission’s discussions around the tokenized US Treasuries market that reached roughly $7 billion in late 2025 have already raised the question of how disclosure rules apply when the asset itself sits on a privacy-preserving network. The Office of the Comptroller of the Currency has, since 2021, signalled that national banks can engage with public blockchains under existing risk-management frameworks, and the door it opened then has stayed open.
The Commodity Futures Trading Commission’s technology advisory committee has held sessions on ZK applications in derivatives reporting. None of this constitutes formal rule-making, but the cumulative signal is clear: US regulators are not going to ban privacy cryptography in finance. They are going to ask what the audit interface looks like, who holds the keys, and how a court order is satisfied. Vendors that answer those three questions early will move fastest.
The Financial Crimes Enforcement Network sits at the awkward centre of the regulatory question. Its remit is to receive Suspicious Activity Reports and Currency Transaction Reports at scale, and the entire framework assumes a bank can identify the parties to a transaction. ZK does not erase that assumption, but it changes who sees what, when. The early FinCEN signal has been cautious engagement rather than open hostility, which is the most that any privacy-tech vendor can currently ask for.
What banks and fintechs are doing with it
The first commercial use cases are not the glamorous ones. KYC sharing across institutions, sanction-screening attestations and selective-disclosure proofs for accredited-investor checks are where the early wins sit. Polygon ID and similar identity stacks let a wallet prove it belongs to a US accredited investor without disclosing the holder’s name or address, which closes a real friction point for regulated DeFi access. Chainlink’s CCIP work on cross-chain messaging has folded ZK proofs into its security model, and several US custodians have run private pilots on the rails it provides.
None of this is mass adoption. The cost of writing a custom circuit still requires specialised engineers, prover infrastructure is not yet a commodity, and most regulators have not formally signed off on ZK as an acceptable compliance interface. The honest read on 2025 is that the technology has moved from research demos into production-grade pilots at large institutions, while the legal and operational scaffolding catches up. The bridge from pilot to mainstream rail will run through the same compliance and supervisory channels that the digital euro and what it means for crypto and banks is travelling through in Europe, and through the same fintech-versus-incumbent dynamics behind why fintech needs to build for federal employee benefits.
At the application layer, the wins are quietly stacking. JPMorgan’s Onyx team has demonstrated ZK-enabled collateral mobility. State Street has run privacy-preserving fund-NAV publication tests. A handful of Tier 2 banks have used ZK to share fraud signals with consortium partners without disclosing the customer file. Each pilot looks like a small experiment in isolation. Stitched together, they trace the outline of a compliance layer that does not require a single shared database to function.
The vendor stack is also consolidating. Aleo, Aztec, RISC Zero and StarkWare each own a slightly different slice of the developer experience. Cloud providers are watching the market for prover acceleration as a managed service. The next twelve months will likely see one or two of the cloud hyperscalers offer ZK proving as a standard primitive alongside HSMs and key management, which is the point at which integration cost stops being the gating factor for mid-tier US banks.
| Milestone | Date | Primary source |
|---|---|---|
| Goldwasser-Micali-Rackoff define zero-knowledge proofs | 1985 | STOC ’85 conference paper |
| Zcash launches first production zk-SNARK blockchain | October 2016 | Zcash Foundation announcements |
| NIST finalises first post-quantum cryptography standards (FIPS 203/204/205) | August 2024 | NIST press release |
| OCC Interpretive Letter 1183 rescinds 2021 non-objection requirement for bank crypto activity | March 2025 | OCC IL 1183 |
| GENIUS Act stablecoin framework signed into US law | July 2025 | Latham & Watkins summary |
What comes next
The next eighteen months in US ZK finance will be defined by three forces. Hardware acceleration is pushing prover times into the millisecond range, which makes real-time use cases viable. Regulatory dialogue is moving from informal sessions to draft guidance, particularly around digital identity and tokenised securities. And the rollup wars on Ethereum are forcing rapid iteration on developer tooling, which lowers the integration cost for non-crypto-native institutions.
None of that means ZK will eat compliance overnight. It means that, by the end of 2026, a US bank running a sanctions-screening attestation through a zero-knowledge proof will look unremarkable rather than experimental. The interesting question is which infrastructure providers end up sitting between the regulator and the rail, and whether they will be crypto-native firms that have learned compliance or incumbent vendors that have learned cryptography.
