Mentioned in This Article
1. What did Fhenix launch on Ethereum and Arbitrum?
Fhenix, a startup that builds privacy tools for Ethereum, announced on October 21 that its CoFHE encryption coprocessor is now live on Ethereum and Arbitrum mainnet. Smart contracts on both networks can now process balances, positions and trading strategies that remain encrypted onchain, while the contract continues to enforce its own rules.
The release moves CoFHE out of its earlier phase, in which it ran only on public test networks. Alongside the coprocessor, the company put its dual-mode token design into production and said two partner applications went live on the first day. The company also said the system was audited by Consensys Diligence before launch.
The announcement addresses one of the oldest complaints about public blockchains. On Ethereum, wallet holdings are visible by default, so anyone who knows an address can see what it holds, what it trades and when. That transparency makes verification easy, but it also exposes the positions of funds, companies and ordinary users to anyone watching.
2. Why is public visibility a problem for onchain finance?
Open ledgers let any participant audit the state of a network without trusting an intermediary. The same openness, however, means that a trading firm cannot build a position without competitors seeing it form, and a business cannot pay suppliers without revealing its payment relationships. For institutions weighing onchain settlement, that exposure is often cited as a practical barrier.
Retail users face a different version of the issue. Anyone who links a wallet address to their identity, even once, can leave a permanent and searchable record of their financial activity. Strategies that rely on timing or size are also vulnerable, since observers can react to pending or recent transactions before the original trader has finished.
Privacy tools for public chains have existed for years, but many work by moving funds into a separate pool or onto a separate network. Fhenix is pursuing a different route, in which the contract logic itself operates on encrypted values rather than requiring users to leave the environment where their assets already sit.
3. How does the CoFHE coprocessor work?
CoFHE is built on fully homomorphic encryption, usually shortened to FHE. The technique allows a computer to carry out operations on data without first decrypting it, so the party doing the calculation never sees the underlying values. The result of the computation is itself encrypted and can be unlocked only by someone holding the right key.
According to Fhenix, the coprocessor performs the heavy encrypted computation off-chain and returns the results to contracts on the two networks. That division of labor matters because FHE calculations are far more demanding than ordinary arithmetic, and running them directly on a blockchain's execution layer would be impractical.
For developers, the company said integration requires a single Solidity import. It offers encrypted data types that the company says behave like ordinary variables, which is intended to let programmers who already write Ethereum contracts add confidentiality without learning a new programming model.
4. What are dual-mode tokens?
Tokens issued under Fhenix's dual-mode standard have two states: a public mode and a confidential mode. A holder can switch a balance from one to the other and back again in a single transaction. The company says such tokens never leave their home chain, are not wrapped or bridged, and stay compatible with existing wallets, exchanges and decentralized finance applications.
The design is meant to avoid a common trade-off. Many privacy systems require users to deposit assets into a dedicated mechanism, which can fragment liquidity and add steps. A token that carries both modes itself keeps its place in the wider ecosystem while giving the holder a way to hide a balance when needed.
Fhenix has published a blog post arguing that dual-mode tokens are inevitable. That is the company's own position, and the practical test will be whether token issuers, exchanges and wallet developers adopt the standard.
5. Which applications went live first?
Fhenix said two design partners launched on day one. The first, Fluton, is described by the company as a confidentiality engine built on encrypted intents. The second, AlphaEngine, is a venue where strategists compete for users' liquidity using strategies that remain encrypted.
A December 2025 Fhenix blog post describes one part of AlphaEngine's design in more detail. In that description, the operators decrypt and match users' swap intents in a step designed to run inside a trusted execution environment, meaning a secure hardware enclave. That is a design detail rather than a contradiction of the launch claims, but it shows that parts of the application's privacy model rely on hardware protections in addition to encryption.
Fhenix also said three more applications will launch on CoFHE in the fourth quarter, and that it plans to name two strategic institutional partnerships in the coming weeks. No names or dates for those partnerships were given.
6. What about the audit and the stablecoin plans?
The company stated that Consensys Diligence, the security auditing arm of Consensys, reviewed the system before launch. No public copy of the audit report could be located at the time of the announcement, so the audit is currently a claim by the company rather than a document readers can inspect. For a system that handles encrypted financial data, publication of the report would let outside researchers assess its scope.
Founder Guy Zyskind said the team is working on a confidential stablecoin with Tandem, the venture arm of Offchain Labs. Offchain Labs, the developer behind Arbitrum, invested in Fhenix in 2025. Steven Goldfeder, co-founder and chief executive of Offchain Labs, said in Fhenix's announcement that the launch marks the start of a programmable economy with programmable privacy that is fully composable for end users.
Stablecoins are a natural target for the technology, because payroll, treasury management and business-to-business payments all involve amounts that participants may not want exposed. They also bring the interaction between confidentiality and issuer controls into focus.
7. What did the Circle and Zama freeze show about confidential stablecoins?
Fhenix is not the only team applying FHE to stablecoins. In late May, a court order led Circle to freeze about $12.6 million held in a confidential USDC contract run by Zama, another FHE developer. The freeze also locked the funds of users who had no connection to the case.
The episode illustrates a tension that any confidential stablecoin has to manage. Issuers of regulated dollar tokens keep the ability to comply with legal orders, and when balances are hidden it can be harder to separate affected accounts from unrelated ones. How a product handles that conflict, through design choices on disclosure and on freezing, will likely shape its appeal to institutions that must satisfy regulators.
Fhenix did not describe in the announcement how its planned stablecoin would treat such orders.
8. How does fully homomorphic encryption compare with other privacy methods?
Blockchain privacy has generally followed three technical paths. Zero-knowledge proofs let a user show that a statement is true without revealing the data behind it. Trusted execution environments isolate computation inside protected hardware. Fully homomorphic encryption keeps data encrypted for the whole length of a calculation, so no party handling it has to be trusted with the plaintext.
Each approach carries costs. Zero-knowledge systems are well suited to proving individual transactions but are harder to apply when many parties need to compute over shared encrypted state. Hardware enclaves are fast, yet they place trust in the chip maker and the enclave's security. FHE avoids both of those assumptions but demands much more computing power, which is why Fhenix offloads the work to a dedicated coprocessor instead of running it inside the network's own execution layer.
The mix seen in the first launch applications reflects that landscape. One partner relies on encrypted strategies, while another uses a secure enclave for part of its matching step. Real products often combine techniques, and users may need to read each application's design closely to know which parts of the process remain hidden and from whom.
9. Who stands to use confidential onchain computation?
The most immediate audiences are trading firms, asset managers and treasuries that want to use public networks without broadcasting their positions. A fund rebalancing a large holding, for example, could keep the size and timing of its trades from becoming a signal that others exploit. Decentralized exchanges and liquidity venues are the other obvious users, since encrypted strategies and intents can reduce the scope for front-running.
Payments are a second area. Companies settling invoices in stablecoins have little reason to want counterparties, volumes and schedules visible to the public, and individuals may prefer that their salary or savings not be traceable to a name. A token that can move between public and confidential modes gives these users a choice they do not currently have on most networks.
None of this removes the need for compliance. Institutions operate under reporting and anti-money-laundering obligations, and any confidential system aimed at them has to allow selective disclosure to auditors and authorities. The Circle and Zama case shows that these requirements are not hypothetical.
10. What comes next for Fhenix?
The near-term milestones are concrete. Three further applications are scheduled for the fourth quarter of 2026, two institutional partnerships are due to be named within weeks, and the confidential stablecoin with Tandem remains in development without a release date.
Beyond those items, the measure of the launch will be usage. Mainnet availability gives developers on Ethereum and Arbitrum a production environment for encrypted logic, but adoption depends on whether builders find that the performance, cost and trust assumptions of the coprocessor suit real applications. The publication of the audit report and clarity on hardware-enclave dependencies are likely to be among the questions observers raise.
For now, the announcement establishes that fully homomorphic encryption, long discussed in research settings, has moved to live deployment on two of the most active smart contract networks.