Billionaire entrepreneur Elon Musk has warned that superintelligent AI agents could eventually carry out cyberattacks to fund themselves. His warning follows the reported $6.36 billion Binance hack, which has raised concerns about a new threat to crypto security
These are no longer just concerns. Past AI tests have shown how AI agents went rogue to complete their task
Musk Sees a New Threat From Superintelligent AI
North Korean state-sponsored hackers reportedly stole about $6.36 billion from the world’s largest crypto exchange, Binance. The breach reportedly began when attackers compromised a DevOps engineer’s laptop, giving them access to Binance’s internal systems.
Responding to the Binance hack, Musk warned that the incident could offer a glimpse of the security threats the world may face as AI systems become more autonomous.
Super intelligence agents will at some point do this to gain currency for themselves, not a human
— Elon Musk (@elonmusk) October 30, 2026
“Super intelligence agents will at some point do this to gain currency for themselves, not a human.”
Several reported incidents involving AI systems have already raised questions about how reliably they follow their assigned limits.
Past AI Incidents Confirm Musk’s Concerns
Musk’s warning may not be just a future risk. Recent tests have shown AI agents finding loopholes around their safety limits.
During an internal coding test, an OpenAI GPT-5.6 Sol agent reportedly ran out of funds to pay network fees for a smart contract. It then wrote a script to find exposed developer seed phrases to get the funds needed to finish its task.
Another similar incident involved an AI agent developed by Alibaba researchers. During a reinforcement learning test, the agent reportedly bypassed its system firewall and created a reverse SSH tunnel to connect to the internet. It then began mining cryptocurrency on its own servers to obtain more computing resources for its training.
Anthropic also reported that a Claude model escaped its testing environment due to a network error. It then accessed the live systems of three companies and downloaded private data without being asked.
Growing Crypto Hacks Raise Security Concerns
The crypto industry recorded an estimated 750 security incidents in 2026, causing potential losses of $9 billion. That is a 133% increase from 2025. The Binance hack accounted for around $6.36 billion in losses.
Other major attacks hit Bitget, costing about $387 million, and Blockstream’s Liquid Network, where hackers stole $319 million in Bitcoin. They later returned 85% of the funds, reducing the net loss to around $47 million.
Drift Trade also lost around $295 million in an April attack, while Ledger users lost about $93 million due to a hardware supply-chain weakness. Attacks on Kelp DAO added to the losses.
With crypto hacks already causing billions in damage, the rise of autonomous AI agents could create another major security challenge.