Our public testnet is running a four-validator, 1-second Aura slot configuration with GRANDPA finality. We are collecting hash-chained audit evidence for a 72-hour acceptance review that began on 22 September at 11:45 UTC. The review is not yet complete, and we have not declared the network accepted.
The remaining gates include fee-failure and refund tests, a controlled P2P disconnect-and-recovery exercise, and a full review of finality, timing, storage growth, and service events. A separate 800 ms lab profile is a future experiment—not a measured speed of the current public testnet.
The current whitepaper presents Helixium's proof-native Layer-1 architecture, A–T–C–G evidence lifecycle, validator model, planned 46 million HLX allocation, and the boundaries between implemented testnet behavior and proposed mainnet policy.
It is a pre-mainnet technical document. Planned economic controls and the maximum-supply lock are not represented as already immutable on the live testnet.
Across four validators, the 24-hour storage test recorded 134.4–195.3 MiB of active database growth per validator. No missed slots or node and lifecycle-worker restarts were recorded. The earlier unbounded RocksDB obsolete-WAL growth pattern did not recur in the active ParityDB paths.
One temporary NTP synchronization alert was investigated and recovered without observed block, finality, peer, or slot impact. This storage result does not substitute for the separate 72-hour distributed acceptance review.
