Can Miden, which has spun out of Polygon's incubation ecosystem, actually deliver on the unmet demand for privacy finance?

Written by KarenZ, Foresight News

After more than five years of research and development, Miden is now being tested with real user capital.

On October 26, the project, which was incubated by Polygon before spinning out into an independent entity, launched its partner mainnet, allowing partners and invited users to test its privacy-first redesigned blockchain with real assets.

A Long Technical Journey Spun Out of Polygon

Miden is no new project. Polygon first unveiled Polygon Miden and an early prototype of the Miden VM back in 2021. At the time, it was part of Polygon's broader zero-knowledge technology roadmap, focused on exploring Ethereum scalability via STARK proof technology: enabling validators to confirm that a program was executed correctly by verifying cryptographic proofs, rather than re-running the full underlying computation that the proof attests to.

This roadmap is directly tied to the background of the project's technical founder, Bobbin Threadbare. He previously served as a lead zero-knowledge researcher at Facebook, and is a core contributor to both the Distaff VM and Winterfell. The two projects serve distinct purposes: Distaff VM explores how to generate STARK proofs for program execution, while Winterfell is a general-purpose toolset for proof generation and verification, built by Facebook's Novi Research team to lower the barrier for developers adopting this class of cryptographic technology.

The early Miden VM was built directly on top of this prior work: it evolved from the Distaff VM, and integrated Winterfell to replace its underlying proof system.

In April 2025, Miden spun out of Polygon's incubation ecosystem, announcing a $25 million seed round led by a16z crypto, 1kx, and Hack VC. The project is co-led by Bobbin Threadbare, Dominik Schmid, and Azeem Khan. It positions itself as an "Edge Blockchain": shifting a portion of the computation and data management that would traditionally be handled by the network onto user devices.

How Does Miden Work?

Miden splits the process of transaction execution, proof generation, and network validation into distinct, modular layers.

In the local transaction model, a user's device first executes the transaction, then generates a STARK zero-knowledge proof that is submitted to the network for validation. The proof attests that the change to account state complies with protocol rules and the account's embedded code. Users can also delegate proof generation to a remote service, though they will need to account for the scope of data disclosed to the service provider.

The entire system operates around two core primitives: accounts and Notes. An account can hold assets, store data, and execute code, and every account is itself a smart contract. The state of public accounts is stored on-chain; the full state of private accounts remains off-chain, with only the cryptographic commitment to the state recorded on-chain for verification. Whether an account is public or private is set at the time of creation, encoded into the account ID, and cannot be toggled for a given account afterward.

Notes serve as the carrier for transferring assets and instructions between accounts. A Note can carry attached assets, and also embed program logic that defines how those assets can be claimed or used. As such, a Note is not merely a simple transfer message, but can also carry executable operations that only run when specific predefined conditions are met.

A typical targeted transfer is split into two steps: the sender first creates a Note that carries the transferred assets, and restricts claims exclusively to the designated recipient account; the recipient then consumes that Note via a separate transaction to credit the assets to their account. For this reason, what users perceive as a single payment in daily usage may map to two separate transactions on the Miden protocol.

Miden will also support network transactions, where the network itself executes transactions and generates proofs, primarily to handle shared contract state used across multiple users. Network transaction builders monitor relevant Notes, and process requests in alignment with the rules defined by the target account.

The public/private attributes of accounts and Notes exist as two independent dimensions. A private account can create a public Note, and making that Note public does not convert the associated private account into a public account. Even when using a private Note, metadata such as the sending account ID, Note type, and tags remain public — "privacy" does not mean there is zero observable information on-chain.

This design draws inspiration from the Actor Model: each account manages its own internal state, collaborates with others via messages, and operations with no interdependencies can be processed in parallel. Transactions that have sequential dependencies must still be executed in the required order.

From Bread to Private Multisig: Privacy Needs to Be Usable First

The user entry point is Bread, a self-custodial wallet available on iOS, Android, and Chrome. Invited users can send and receive payments, swap assets, and perform cross-chain transfers; eligible users can also deposit funds directly from the wallet into Aave on Ethereum.

Transaction fees are paid in USDCx. This asset is issued by Miden via Circle xReserve, and is 1:1 backed by USDC held in the xReserve contract. The pilot wallet also comes pre-loaded with a small amount of fee-paying assets, to cut down the friction of users needing to source fee tokens before they can start interacting with the network.

Since private data is stored off-chain, account recovery cannot rely solely on public on-chain records. Guardian, an auxiliary service co-designed by Miden and OpenZeppelin, provides state backup, sync, and recovery capabilities. The service's current operator can view account state, but cannot independently move funds with only its own signature.

For enterprise and team use cases, accounts require shared management by multiple parties. Miden has already launched private multisig on its partner mainnet: participants sign independently off-chain, and an operation can only be executed once the predefined user approval threshold is met, along with any other authorization requirements set for the account. Teams can also require additional signatures for larger transactions, without publishing balance, signer structure, approval threshold, or workflow details as public on-chain data.

This set of products is built to address the demand for retaining internal approval, recovery, and control capabilities alongside financial confidentiality. The Miden team has also published a list of applications built on Miden, including Qash, a B2B neobank focused on payroll and treasury management; Sempo, which builds borderless payment and digital dollar account infrastructure; and Darwin, which is exploring tokenized equities and structured investment products.

Real Funds Are Here, But Privacy Still Has Clear Boundaries

The Miden partner mainnet is an invite-only, temporary pilot stage. The team requires all participants to withdraw their funds ahead of a planned future network reset, with withdrawal processes and deadlines to be announced separately. Both the Bread wallet and cross-chain interface implement deposit limits to cap the total balance in the Bread wallet at under $1,232.462; this unusual number is a nod from the team to the underlying Goldilocks prime number used in Miden's cryptography.

Bobbin outlined two key privacy guardrails in the current design: validators have access to private transaction inputs to perform additional proof verification; transaction details are archived using threshold encryption, which can only be decrypted by a predefined quorum of validators in the event of an anomaly. As a result, private inputs are not disclosed to the general public, but are not kept secret from validators — and this applies even to private multisig transactions.

In addition, swaps are currently fully public, and cross-chain and external services will have access to the information required to process related transactions. Miden also allows asset issuers to set freeze, hold, and transfer restrictions. As such, users need to independently assess three separate dimensions: account confidentiality, free transferability of assets, and data visibility for connected service providers.

At present, client-side transaction proofs are fully implemented, but block-level proofs are not yet complete, and block producers are still centralized. Bobbin noted that the team plans to transition to a public mainnet within a few months, with planned improvements to fee mechanisms and proof generation.

On the performance front, Bobbin stated that proof generation for most transactions can be completed on a mobile device in under three seconds.

Summary

Token economics have also been mentioned in official disclosures. Polygon noted in its April 2025 announcement that Miden would launch a future native token, with roughly 10% of the supply expected to be airdropped to POL stakers, to incentivize POL stakers to support the Agglayer. In November of the same year, Miden announced the launch of its leaderboard on Kaito, and teased upcoming reward details. However, none of these materials provided a specific token launch timeline or full allocation rules.

The launch of the partner mainnet marks the start of Miden's testing of its technology and products in a real-fund environment. Going forward, the most important metrics to watch are sustained user adoption, growth in real-world business demand, and the stable performance of core features — including privacy protection, multi-party approval, account recovery, and cross-chain functionality — in daily usage scenarios.