We are building the category of verifiable data.
zkDB is an independent research and engagement firm. We design and deliver zero-knowledge database systems for institutions that cannot show their data and cannot afford to be doubted — in regulated finance, healthcare, government, and critical infrastructure.
We take a small number of deep engagements at a time, we publish the conceptual scaffolding of the field, and we do not productize.
A decade of peer-reviewed verifiable-database research behind the practice
Regulated sectors the firm is built for — finance, healthcare, government, infrastructure
Compliance frameworks held in scope across engagements
Client rows our reference architecture requires the firm to see
For the entire history of computing, to be convinced of a claim you had to see the data behind it.
That assumption is now obsolete. Zero-knowledge proofs make it possible to verify that a query was run correctly on real data without that data ever changing hands. We believe this is a once-in-a-generation shift in how institutions exchange trust — on the order of public-key cryptography — and that within a decade the ability to prove a result without revealing it will be a baseline expectation of regulated data infrastructure.
The firm exists to bring that shift to the institutions where it matters most, and to do so before it becomes obvious. The mathematics is published and peer-reviewed; the hard, defensible work is the engineering of fitting it to a specific institution’s data, regulator, and risk perimeter. That is our discipline.
The cryptography comes from the blockchain world. The architecture deliberately does not.
The hard machinery behind a zero-knowledge database — PLONKish arithmetization, Halo2, KZG and IPA polynomial commitments, recursive proof composition — was forged and battle-tested in the public blockchain and zk-rollup ecosystems, where it secures tens of billions of dollars under relentlessly adversarial conditions. That is the most demanding proving ground cryptography has ever had. We inherit its maturity — and leave the rest of the chain behind.
- Proof systems hardened in production — Halo2, PLONKish circuits, KZG & IPA commitments, recursion.
- A decade of open, peer-reviewed cryptanalysis and real-world exposure no enterprise lab could replicate alone.
- Succinct, non-interactive, infinitely re-verifiable proofs as a transferable artifact.
- No chain, no token, no consensus protocol, no gas — and no crypto-asset exposure of any kind.
- Your data, your schemas, and your very participation never appear on any public ledger.
- Integration with your existing data warehouse and governance, not a parallel decentralized stack.
Where an institution wants tamper-evident, independently-anchored commitments, a zkDB can optionally publish its short cryptographic fingerprints to a ledger of the institution’s choosing — public or private. That is a deliberate policy decision, never an architectural requirement.
The principles the firm is built on.
Applied cryptographic research
Our principals work at the seam of production cryptographic engineering, regulated-data systems, and the peer-reviewed literature on verifiable computation. Each engagement is treated as bespoke applied research — never the deployment of a fixed product.
Engagements over sales
No pricing pages. No free tier. No sales funnel. The first conversation is a confidential briefing, and most briefings do not become engagements — because we are explicit about when zero-knowledge proofs are the wrong tool for the problem.
Owned by no ecosystem
No chain, no token, no cloud-vendor or hardware affiliation, no outside commercial pressure on our recommendations. We select proof systems and commitment schemes on technical merit for the institution in front of us — and nothing else.
The firm, not the individual
Our research, writing, and briefings are conducted in the name of the firm; the category is larger than any one person. Our principals are introduced privately to clients under engagement NDAs.
Built to survive your procurement and legal review.
The institutions we work with run rigorous vendor due diligence. We are structured to clear it — and to hold ourselves to the same standard of verifiability we sell.
Confidentiality first
Every engagement opens under a mutual NDA. We never name a client, a sector specific to a client, or the substance of an engagement without written consent. The firm’s public voice and its client work are kept strictly apart.
You own the deliverables
Architecture documents, circuits, provers, verifiers, and governance instrumentation produced for you are yours. We retain only the general, non-confidential methods that make up our practice — never your data, schemas, or domain logic.
Designed to need no raw data
The architectures we build are precisely those in which your sensitive rows never have to leave your perimeter — including ours. Wherever feasible, an engagement is scoped so the firm itself works against synthetic or committed data, not the real thing.
Honest go / no-go
If zkDB is the wrong tool — if a trusted execution environment, multi-party computation, or plain access control fits your problem better — you will hear us say so in the first session, before any commitment. Our reputation depends on it more than any single fee does.
We lead the category in public.
A firm that intends to define a category has to be willing to teach it. We maintain an open conceptual library and track the research frontier — so a technical buyer can evaluate us on substance long before a single conversation.
The vocabulary of verifiable data
Plain-English explainers from first principles to architecture — ZK proofs, polynomial commitments, PLONKish custom gates, lookup arguments, and where zkDB sits beside FHE, MPC, TEEs, and differential privacy.
The frontier we build on
The peer-reviewed work — across a decade of verifiable-database research — that our engagements stand on, read for what it means in production rather than in the abstract.
Where it changes everything
The regulated domains the practice is built for, and the specific institutional problems where proving a result without revealing the data is the unlock.
For journalists and editors.
We are glad to provide background, context, and on-the-record commentary on the state of zero-knowledge data systems, the regulatory landscape, and adjacent privacy-enhancing technologies. We do not comment on specific clients or engagements.
Bring us your hardest data trust problem.
Briefings are confidential, deeply technical conversations with our principals. We work with regulated enterprises, central institutions, and serious research programs.