Automated Management via Stakenet

How JitoSOL stake moves automatically

Over time, as people deposit SOL into the reserve, or as validator performance varies, it becomes necessary to move stake around. Jito now uses Stakenet – a blend of on-chain programs and off-chain "keepers" – to automatically manage JitoSOL's stake distribution across validators.

Design Principles

On-chain decision making
The Steward and Validator History programs manage stake allocation decisions directly on Solana, ensuring transparency and verifiability.

Community governance
All system parameters can be adjusted through Jito DAO governance, allowing the community to influence delegation criteria and safety mechanisms.

Decentralized execution
Multiple independent keepers execute the system's decisions, removing single points of failure while maintaining network health and optimal yields.

Evolution from the previous stake-bot

The original Kobe system was an automated off-chain service that stored network information about validators and delegated stake according to each validator's performance. While functional, this approach had limitations:

Centralized signing authority
Off-chain logic that was harder to audit
Limited community visibility into decision-making

Stakenet addresses these limitations with an on-chain-first design where logic is transparent and keepers are permissionless.

Stakenet architecture

Component	Function	Location
Validator History Program	Stores up to 512 epochs of validator performance, commission, and MEV data	On-chain program
Steward Program	Evaluates validators, determines target stake allocations, and manages rebalancing via a state machine	On-chain program
Keeper Bot	Executes the permissionless transactions required by the Steward and uploads validator data	Off-chain service (Docker or binary)

The system operates with on-chain programs making decisions and off-chain keepers executing those decisions.

How stake rebalancing works

Scoring phase – Every 10 epochs, the Steward evaluates over 1,000 validators
Target calculation – The top 200 validators (configurable) are selected, each receiving an equal target allocation
Monitoring – Keepers continuously update validator performance data
Instant unstake evaluation – Each epoch, validators that meet certain negative criteria are marked for immediate removal from the pool
Rebalancing – Keepers execute stake delegation and undelegation transactions to move validators toward their targets, subject to safety caps
Cycle repetition – The process continues until the next scoring phase

All parameters governing this process are stored on-chain and can be modified through Jito DAO governance.