Bybit has launched AI sub-accounts, a new account type designed to let AI trading agents operate with dedicated fund isolation and permission controls, the exchange announced.
What Bybit’s AI sub-accounts are built to do
The new AI sub-account structure allows users to delegate trading access to AI agents without exposing their main account funds. According to Bybit’s official announcement, the feature pairs fund isolation with granular permission controls, meaning an AI agent can only access the capital and trading functions explicitly assigned to it.
The sub-account model separates AI-driven activity from a user’s primary holdings. If an AI agent malfunctions or executes poorly, losses are contained within the sub-account’s allocated balance rather than affecting the trader’s full portfolio.
Bybit’s help center documentation outlines the setup process, confirming that AI sub-accounts operate as distinct entities under the main account with their own permissions and fund pools.
KEY POINTS
- Fund isolation: AI agents trade only with capital explicitly allocated to the sub-account
- Permission controls: Users set granular access limits on what actions AI agents can perform
- Account separation: AI trading activity is walled off from the main account balance
Why risk controls are central to AI trading agent adoption
As exchanges begin accommodating AI-driven trading, the core challenge is trust. Automated agents make decisions without human confirmation on each trade, which creates operational risk if access is too broad or fund boundaries are unclear.
Bybit’s approach addresses this by treating AI agents as restricted sub-users rather than giving them full account access. The permission layer lets traders define exactly which markets, order types, or position sizes the agent can use. The move comes as major exchanges expand their platform offerings to serve increasingly diverse user bases, from regional retail traders to automated systems.
The fund isolation component acts as a containment mechanism. Rather than relying solely on an AI agent’s internal risk logic, the exchange enforces hard limits at the infrastructure level. This is a meaningful distinction: the risk controls exist at the platform layer, not just within the agent’s own code.
Infrastructure upgrades like these are part of a wider industry push to support institutional-grade tooling, a trend also visible in events such as the GovXcellence Summit Malaysia 2026 where digital infrastructure governance is a central theme. Similarly, the growing demand for data center capacity to power AI and crypto workloads is drawing attention at venues like the World Datacentre Summit Vietnam 2026.
This launch signals that exchange-level account architecture is adapting to a future where AI agents are routine trading participants, with scoped permissions and fund isolation likely becoming baseline requirements across the industry.
Disclaimer: This article is for informational purposes only and does not constitute financial or investment advice. Cryptocurrency and digital asset markets carry significant risk. Always do your own research before making decisions.
