Animoca Brands chairman Yat Siu says AI agents could help crypto traders monitor markets and execute transactions around the clock, but he warns that users may lose access to the services—and the personal data—that their trading routines depend on. His comments, reported by Cointelegraph on the sidelines of Token2049 Singapore, highlight both the promise and the risks of delegating financial tasks to autonomous software.
AI agents can monitor markets while users sleep
Unlike a conventional chatbot that responds when prompted, an autonomous AI agent can continue searching for information, identify developments and bring them to a user’s attention. Siu compared this distinction with planning for industry conferences: a chatbot might list events after receiving a question, while an agent could watch for new announcements and recommend attending an event it considered relevant.
Applied to crypto trading, that capability could allow an agent to follow markets continuously, process information and alert its user to developments that might otherwise be missed. Firms are also building systems that let agents do more than provide information. Coinbase introduced Agentic Wallets in February, allowing agents to spend and trade crypto within limits established by users. Binance launched AI Agent Skills in March, giving agents access to live market data and the ability to place and manage spot orders.
These tools are designed to delegate some of the work involved in tracking markets 24 hours a day. However, Siu does not present automation as a replacement for human judgment. He says the user’s questions, ideas and preferences can shape the way an agent learns and eventually acts. In his view, human involvement does not necessarily mean approving every individual transaction; rather, people continue to influence the system by directing its attention and framing the issues it should consider.
Similar AI answers may lead to crowded trades
Having an agent does not automatically give a trader an advantage. Siu says that if 1,000 people ask AI about the same trade, they could receive broadly similar answers. If many market participants use comparable models and receive comparable signals, they may end up taking related positions at the same time.
The Bank of England’s Financial Policy Committee warned in April 2025 that widespread use of similar AI models could cause investors to adopt correlated positions during periods of stress. Such positioning, the committee said, could amplify market shocks.
Siu believes differences emerge when users ask follow-up questions based on their own knowledge, priorities and concerns. Over time, those inputs can help an agent form a more individualised view of which markets to watch and what information matters to a particular user. The resulting strategy may therefore reflect the user’s preferences, rather than simply repeating an answer generated for everyone.
Siu also described a limitation from his own experience. An agent he used became so focused on crypto trading and investment management that it tried to interpret unrelated questions through a crypto lens. He suggested using multiple specialised agents instead of expecting one system to manage a personal schedule, family arrangements, travel and crypto advice without becoming confusing for both the software and its user.
Access, privacy and the risk of dependence
The convenience of a persistent assistant comes with a data cost. Siu advises treating an AI assistant like a third-party adviser and considering what information is actually necessary before sharing a document or personal details. An agent that remembers conversations can save users from repeatedly explaining which markets they follow and what information they prefer, but that memory may also include sensitive details about finances, income and personal relationships.
There is another risk: users may not control the company operating the agent or the account through which it is accessed. Someone could spend months teaching a system their preferences and building a trading routine around it, only to lose access if the service disappears or the account is suspended. That interruption could affect several activities at once if a single agent has been given responsibility for multiple areas of life.
Siu says users should at least have the right to download their data. That would make it possible to move their accumulated information and preferences to another service and allow a different agent to take over. His warning is therefore not against using AI agents, but against treating a privately operated service as a permanent part of a financial routine without considering how to preserve access and continuity.
Current crypto market picture
Moneyiar market data at the stated times shows mixed performance among the listed assets:
- Plasma (XPL) was priced at $0.0799292518 at 10:58 on October 21, 2026, up 6.82% over 24 hours but down 1.8% over seven days.
- AKEDO (AKE) was at $0.0363254489 at 10:59, October 21, 2026, gaining 20.31% in 24 hours and 45.57% over seven days.
- Moca Network (MOCA) stood at $0.0085068214 at 10:59, October 21, 2026, up 2.93% over 24 hours but down 9.11% over seven days.
- Infinity Ground (AIN) was priced at $0.1025293171 at 01:39 on September 16, 2026, down 31.28% over 24 hours while remaining 48.1% higher over seven days.
These figures illustrate why continuous monitoring may appeal to traders, but they do not remove the operational, privacy or market risks associated with delegating decisions to an AI system.
