Skip to main content
A Meta Vault separates deciding what the portfolio should look like from actually moving capital to match. Weights express the target; rebalance() enforces it. A vault can be off-target between the two steps, and that is expected.

The target allocation model

A Meta Vault maintains a whitelisted strategy set, a target weight per strategy in basis points, and a target idle weight (the share of the portfolio meant to stay undeployed). With an idle weight of 500 (5%), the vault targets roughly 5% idle and 95% deployed across its strategies according to their weights.

Setting targets

Updating weights is a separate action from moving capital: it only changes what the vault is aiming for, not what it currently holds. Weights can be set manually by the vault’s operator, or supplied by yield.xyz’s optional allocation optimizer described in Architecture.
Updating weights does not move funds by itself. Capital only moves when a rebalance runs.

Enforcing targets

A portfolio-wide rebalance compares the vault’s current holdings against its targets, strategy by strategy and for idle, then withdraws from over-allocated positions and deploys into under-allocated ones. This is the step that actually shifts capital, whether it follows a weight update, a strategy being added or removed, or idle capital building up from compounding. A vault operator can also move a specific amount into or out of a single strategy directly, which is how newly deposited idle capital gets deployed for the first time, or how a specific position gets pulled back to idle outside of a full portfolio rebalance.

When automated rebalancing acts

Where yield.xyz’s optimizer is enabled for a deployment, it does not rebalance on every price tick. It compares the drift between current and target weights against a configured threshold and only proposes and executes a rebalance once that threshold is crossed, so the vault avoids churning gas on marginal reallocations. Every proposed and executed rebalance is recorded with its inputs, the algorithm used, and the resulting drift, whether or not automated execution is enabled for that vault.

Next steps

Strategy lifecycle

Add, remove, and replace strategies using these same primitives.

Roles and signing models

Which role can update weights and trigger a rebalance.