What can go wrong.
The machines are not on-chain. Buying hashrate from a provider does not make the provider disappear.
Read these first
- Provider default. A vendor can change terms, delay payouts, or stop sending coins. On-chain code cannot force a warehouse to mine.
- Same-human risk. If the people who write the protocol also sell the hashrate, the “independent protocol” claim is weaker than the copy. Keys, banks and title have to be separate or the claim is false.
- No cash put. You cannot redeem hashrate for dollars. Exit is a market. Markets go quiet.
- Contracts can go dark. When Bitcoin price falls and power does not, net per terahash can hit zero. Some vendor terms let them idle unprofitable machines.
- NAV is an estimate. Replacement cost, observed coins and assumed remaining life can all be wrong. A receipt ages.
- Thin book. Early AMM liquidity can be shallow. Selling a large bag moves the price through the buyback band.
- Oracle and keeper risk. Settlement and SAFE MODE depend on publishers posting expected vs actual. Disagreement freezes mint. It does not recover lost coins.
- Legal uncertainty. A share of a mining vault may be treated as a security in some places. This site is not an offer.
What v1 refuses
Borrowing stablecoins against the Bitcoin treasury to buy more hashrate looks good on a quiet tape. In a crash, collateral and hashprice fall together and a liquidation sells the coins the vault was built to hold. The credit module exists in the spec and stays off.
What “decentralized” means here
Mint authority lives on a program, not a person. Payout addresses are published. Anyone can watch incoming coins. Providers are rows in a registry, not the protocol. The farms themselves stay centralized. Phase two is a second provider. That is the first real cut in vendor risk.