Testnet vs mainnet
Practice funds versus real money: what's identical (everything about how the product behaves) and what's different (the money, the markets available, and the safety rails).
Hyperliquid runs two parallel environments. Testnet is the rehearsal stage: identical mechanics, free mock USDC, zero consequences. Mainnet is the live exchange with real money. Every Hypertrade desk belongs to exactly one of them, chosen at creation and shown on a badge everywhere the desk appears, and you can run desks on both at once.
| Testnet | Mainnet | |
|---|---|---|
| Money | Free mock USDC (faucet or pre-funded practice desk) | Real USDC, deposited by you |
| Markets | Crypto perpetuals | Crypto perpetuals + eligible US-stock HIP-3 perps; commodities (gold, silver, oil, and more) are open for manual trading, while automated commodity strategies stay gated until certified |
| AI + risk engine | Identical | Identical |
| Owner account | Your own, or the managed practice wallet | A user-controlled embedded owner wallet; external registration only when enabled |
| Extra rails | None | Explicit risk acknowledgement + rollout caps on position notional while live execution proves out |
Why testnet first is genuinely worth it
- You learn the interface, the verdict trail, and your own reaction to drawdowns, for free.
- The desk's behavior transfers 1:1: same signals, same rules, same exits. Only fill quality differs slightly (testnet books are thinner).
- It surfaces setup mistakes (wrong network, unfunded perps balance) while they cost nothing. The Check funding doctor works on both networks.
Moving to mainnet
Create a new desk on mainnet (or point an existing one at a mainnet wallet). You'll pass an explicit risk acknowledgement, and the wallet-network truth is enforced: a desk's badge always reflects where its wallet actually lives, so a “testnet” label can never hide real-money trading. Start with an amount you'd be comfortable trading manually at leverage; the engine bounds each day's damage, not the market's direction.