Acurast is a decentralized compute project that uses smartphones to run workloads in hardware-backed Trusted Execution Environments. It has public software, a mainnet launch and published component security reviews. ACU pays for network activity and supports its incentive system; staking exposes locked tokens to compute-performance penalties as well as market risk.
Assessing whether Acurast is legitimate therefore involves two questions: what evidence supports the project, and what obligations come with holding or staking its token?
A project with public infrastructure and disclosed vulnerabilities
Acurast announced its mainnet and token launch on January 20, 2026. Its public chain repository and developer documentation provide material that can be inspected beyond promotional claims.
Security reviews have specific boundaries. Inference’s ERC-20 token assessment reviewed identified contracts and a particular code revision. The scoped review reported no security issues, while recording three open observations. Deployment configuration, associated key management and malicious administrator behavior were outside its scope. It is not a blanket assurance covering every network component or future upgrade.
There is also a disclosed failure to consider. In a September 4 security update, the Association described a component bug that overpaid rewards. It said larger recipients agreed to burn surplus tokens and put the remaining impact at about 71,000 ACU, which it would cover from operational funds. The update described fixes and published post-mortems, attributing the incident to a bug rather than an external network breach.
ACU links compute access, rewards and governance
The documented network model connects developers seeking execution capacity with smartphone processors. Hardware attestation supports verification and confidentiality; users still depend on the software, device security and protocol implementation.
| ACU function | Economic role |
|---|---|
| Network and deployment fees | Supports chain interactions and the scheduling of compute workloads |
| Provider incentives | Rewards compute provision and block production through separate allocations |
| Staked Compute | Places collateral behind a provider’s promised computing capacity |
| Governance | Enables participation in protocol decisions and treasury proposals |
Some developer access routes accept USDC, including the documented Base x402 route. Application use therefore need not mean every customer directly purchases ACU. The investment question is how actual usage, fees and token demand develop alongside incentives and new issuance.
Staking backs a measurable compute commitment
A committer operates phones, locks ACU and promises a level of compute. A delegator contributes tokens to a selected committer without operating hardware. Delegation introduces exposure to that provider’s performance and its delegation fee.
Staking mechanics use four metrics: single-core CPU, multi-core CPU, RAM and storage. Rewards depend on committed compute, stake and cooldown duration relative to other participants. An epoch is 900 blocks, approximately 90 minutes. This is a performance-backed reward allocation, not a fixed interest payment for simply owning ACU.
A longer commitment can increase reward weight, while retaining exposure for longer. A large advertised return is incomplete without the provider’s capacity, fee, recent performance and exit terms.
Slashing can reduce the delegated principal
Under the published slashing rules, compute falling below a commitment in any metric makes that epoch eligible for a penalty. Offline devices or lost capacity can cause a shortfall without malicious conduct. An on-chain slasher triggers the penalty, no earlier than the following epoch.
The documented maximum is 0.003424657534% of stake per epoch. Each affected metric contributes a penalty based on its weight and proportional shortfall. A hypothetical 1,000-ACU stake with a 50% RAM shortfall, using RAM’s 0.4615 weight, incurs about 0.0079 ACU for that metric. Complete failure across all four metrics reaches about 0.0342 ACU for that epoch.
Delegators share the percentage reduction applied to their selected committer’s stake. It affects principal, rather than merely withholding future rewards. Repeated failures can produce repeated penalties; the per-epoch ceiling is not a lifetime loss cap. Redelegating changes future provider exposure but does not undo an incurred penalty.
The exit countdown still carries obligations
The mainnet parameters allow a selected cooldown from approximately 28 to 1,344 days. The countdown starts when exit is requested, not automatically when the stake is created. Mainnet ACU terms should not be confused with the much shorter Canary cACU periods.
The staking FAQ describes reward and vote weights falling to 50% during cooldown while full slashing exposure remains. This is a weight adjustment, not a guaranteed cash-return calculation. Committers must maintain the promised compute until exit is completed. For delegators, a provider’s decision to enter cooldown also matters; the protocol provides redelegation, which is distinct from withdrawing freely transferable tokens.
Inflation rewards participation while expanding supply
ACU tokenomics specifies an initial supply of one billion and a current annual inflation parameter of 5%, subject to governance. One billion is an initial issuance figure, not a fixed maximum supply. The inflation rate is also not an individual staking APR.
Of inflation, 70% goes to Staked Compute, 15% to the treasury, 10% to base benchmark rewards and 5% to collators. Individual rewards depend on competition for those pools. The original allocation also assigns 24% to the team and advisers, with a six-month cliff and 36-month linear vesting.
Issuance and unlocks affect supply through different routes. KTX’s explanation of market capitalization and fully diluted valuation helps distinguish initial, circulating and projected supply. More reward tokens do not ensure a higher dollar value after price changes, fees and penalties.
Base trading has separate bridge and liquidity conditions
The official token registry lists Base ACU at 0xc5fEd7c8cCC75D8A72b601a66DffD7A489073F0b, matching the contract in the KTX Base ACU trading page. It is a bridged instance; native Acurast mainnet ACU has no ERC-20 contract address.
The illustrated quote is about $0.09402, with displayed liquidity near $157,300. Liquidity value is not a promise that a position can be sold at the quoted price. KTX’s liquidity guide explains why order size and price impact matter.
Self-custodied EVM ACU must follow supported bridging routes to native ACU for mainnet staking. Buying a bridged token does not automatically create a stake. KTX’s onchain trading FAQ currently says onchain assets cannot be directly deposited or transferred. Its trading product is therefore separate from funding a self-custodied staking position. Eligible users can register a KTX account to review available trading services.
The useful follow-up evidence is concrete: paid compute activity, fee generation, provider reliability, actual slashing and governance changes. These measures help assess whether the network’s economics support the commitments being made. Project documentation and a correct contract address establish a starting point; they do not set an investment outcome.
This article is educational and does not constitute investment advice. ACU can lose value, and staking, bridging and trading can result in partial or total loss. Numerical examples are illustrative.