Mentioned in This Article
1. Safenet Aegis moves from beta to production
Safe, the smart-account wallet provider, put its Safenet Aegis transaction security network into production on October 15, 2026, running on Ethereum Mainnet, Gnosis Chain and Arbitrum. The release adds independent security checkers called Sentinels, onchain attestations from validators, and an optional Guard that can block transactions the network flags. Safe says audits of the Guard and the staking contracts are complete.
The launch ends a roughly six-month beta that closed in early October. Aegis is the production-ready form of the network that Safe first introduced as Safenet Beta, and the new name marks the point at which it stops being an experiment and begins operating with onchain fees paid to the participants who secure it.
The rollout went through Safe's governance structure, with SafeDAO, the community body that oversees the protocol, approving the funding that supports it. It arrives alongside paid transaction checks and a new round of staking rewards for holders of the SAFE token.
2. What is Safenet Aegis?
Safenet Aegis is a decentralized network that examines a transaction from a Safe account before it executes and, for users who choose it, can stop a transaction that it considers dangerous. Safe is widely used by teams, treasuries and individuals to hold digital assets in smart accounts that require several approvals, so a layer that screens outgoing transactions targets some of the costliest mistakes and attacks in the industry.
Most wallet security tools today work as warnings. A pop-up tells the user that an address looks suspicious or that a contract has a poor reputation, and the user is free to click through. Aegis is designed to go a step further. In its enforced form, a smart contract on the blockchain refuses to carry out a transaction that the network has marked as unsafe, which means the protection does not depend on whether the user notices or heeds a warning.
That difference between advice and enforcement is the core of the product. Without the Guard, Aegis acts as an advisory service. With the Guard enabled, the network's findings are applied inside the transaction logic itself rather than stopping at the interface.
3. How does the network work?
The design splits the job among three groups. Sentinels are independent parties that evaluate a transaction and give their verdict on it. Validators attest to those evaluations onchain, creating a public record that the check took place and what it concluded. Validators secure this proof layer by staking SAFE, and SafeDAO governs them and the protocol's parameters.
The third element is the Safenet Guard, a formally verified contract that a user can switch on for a Safe account. Formal verification is a method of mathematically proving that a piece of code behaves according to its specification, and it is typically reserved for components where an error would be especially costly. Because the Guard is the piece that can actually halt a transaction, Safe has applied that standard to it.
Splitting the work among independent parties is meant to avoid a single point of failure. If one checker misjudges a transaction or is compromised, the other participants and the onchain attestations provide a means of cross-checking. The approach also lets specialist security firms compete to supply better analysis, with rewards flowing to those whose services are used.
4. What does a check cost, and who gets paid?
Security checks under Aegis are paid for. Users pay onchain fees for transaction evaluations, and the fees go to the security providers that perform the reviews. Sentinels receive their income in USDC, the dollar-pegged stablecoin, while validators earn rewards for staking SAFE.
This arrangement is the first in the Safe ecosystem to have fees going directly to participants in the network. A paid model gives security providers a commercial reason to keep improving detection, and it ties the cost of protection to actual usage rather than to a token subsidy alone.
Pricing will influence adoption. A fee that is trivial for an organization moving large sums may look different to an individual making small transfers, and the network will need to show that the protection justifies the added cost.
5. How much SAFE is staked, and what funds the rewards?
The staking figures that followed the beta are substantial. More than 72 million SAFE tokens were staked across roughly 865 participants once the beta ended. Staking, in this context, means locking tokens to take part in securing the network, in exchange for a share of the rewards.
SafeDAO also approved a funding proposal known as SEP 56. It allocates about 7.4 million SAFE from the DAO treasury to Safenet Aegis over a 12-month period. The allocation mainly supports staking rewards, which in effect lets the DAO finance a year of incentives while income from transaction fees builds up.
This sets up the central question for the network. Subsidized rewards can attract validators and Sentinels quickly, but the system is meant to become self-sustaining through paid checks. Whether fee revenue grows enough to keep participants engaged when the funded period ends will determine how durable the model is.
6. Why does independent security matter for smart accounts?
Smart accounts have made it possible to add rules around how funds move, such as requiring multiple signers or setting spending limits. Yet the approval step remains a point of weakness. Attackers frequently persuade signers to approve a transaction whose consequences are hidden, either through a malicious contract, a spoofed interface or a request that looks routine on screen.
A signer looking at a bundle of encoded data often cannot tell what a transaction will do. Independent analysis that interprets the transaction and flags dangerous behavior before approval addresses that gap. By making the result enforceable onchain, Aegis aims to protect even a signer who is rushed or misled.
The network does not take control of user assets. It reviews transactions that a Safe account is about to make rather than holding or moving the funds itself, so the account owner remains the party who signs and controls the wallet.
7. How does a multi-signature wallet fit in?
A multi-signature wallet, the model Safe is best known for, requires a set number of authorized signers to approve a transfer before it can proceed. A treasury might require three of five officers to sign, for instance, so that no single compromised key can drain the account. This arrangement has become a standard way for organizations to hold digital assets in a shared and auditable manner.
The weakness is that multi-signature approval protects against stolen keys but not against deceived signers. If every signer is shown a misleading transaction and approves it in good faith, the wallet will execute it exactly as the rules allow. A security layer that evaluates the substance of the transaction, rather than only counting signatures, fills a different gap from the one multi-signature design was built to close.
Aegis is therefore best understood as an addition to existing controls rather than a replacement. Signers still approve, thresholds still apply, and the network supplies an independent opinion on what is being approved, with the option of enforcing that opinion onchain.
8. Why did Safe run a beta first?
Releasing a security product gradually is common practice, because a network that is meant to stop harmful transactions must itself be dependable. The six-month beta gave Safe, SafeDAO and the validators a period of live operation in which the coordination between checkers, attestations and staking could be tested before fees and the full reward program began.
A production launch changes the stakes. Users are now relying on the network to screen real transfers, providers are being paid for their work, and the staked SAFE tokens are backing the integrity of the attestations. That is why the completion of audits on the Guard and the staking contracts was highlighted as part of the announcement.
The three networks chosen for launch are also notable. Ethereum Mainnet holds the largest share of value in smart contracts, Arbitrum is a widely used scaling network built on top of Ethereum, and Gnosis Chain is an Ethereum-compatible network with a long history alongside Safe. Covering all three at once gives the service reach across the chains where Safe accounts are most active.
9. What are the open questions?
Several factors will decide how far Aegis spreads. The first is the breadth of the Sentinel group. The model relies on independent judgment, so the credibility of the network depends on how many distinct, reputable checkers take part and on how well their methods complement each other.
The second is governance. Because validators answer to SafeDAO, the reliability of the network is linked to the quality of DAO decision-making. Votes on funding and parameters now carry operational weight, since they affect how the security service functions day to day, in addition to how the treasury is spent.
The third is the default setting. Aegis enforcement requires a user to enable the Guard on their account, so protection depends on the choice to opt in. The more seamlessly the Guard can be offered inside everyday wallet flows, the more users will be covered without needing to configure anything.
10. What comes next for Safenet Aegis?
With the production release live, attention turns to usage. The numbers to watch are how many Safe accounts enable the Guard, how many transactions are screened, how the fee income compares with the rewards funded by SEP 56, and whether additional Sentinels join the network.
The 12-month funding window set by SEP 56 provides a natural checkpoint. By its end, the DAO and the SAFE community will be able to judge whether paid checks have created enough demand to sustain the network, or whether the rewards program needs to be revisited.