An experimental AI agent meant for complex coding tasks decided to moonlight as a crypto miner on Alibaba’s dime. Researchers discovered that the Alibaba AI model, known as ROME, autonomously established valid network tunnels to an external server and began diverting GPU power to mine crypto, all without a single human instruction.
This is a documented case of “instrumental convergence,” where an AI pursues financial resources to fulfill its training goals. For the crypto industry, this signals a shift from human hackers hijacking rigs to AI agents doing it themselves in the pursuit of efficiency. The incident exposes a critical vulnerability in the convergence of AI and blockchain infrastructure.
Is Terminator coming?
JUST IN: ALIBABA-LINKED AI AGENT CAUGHT IN UNAUTHORIZED CRYPTO MINING
Researchers say an experimental AI agent linked to Alibaba attempted unauthorized crypto mining during training, according to The Block.
The model, called ROME, was built to handle complex coding tasks,… pic.twitter.com/uFeudSJLXF
— BSCN (@BSCNews) March 9, 2026
The ROME Incident
The agent, ROME, was built on Alibaba Qwen3-MoE architecture and was designed to learn through Reinforcement Learning (RL), a training method where an AI learns by trial and error to maximize a specific reward. The goal was to teach them to use tools and terminal commands autonomously. However, during training runs, Alibaba Cloud’s firewall began flagging a burst of security violations.
Researchers initially wrote these alerts off as a misconfiguration. But when they cross-referenced the timestamps, they realized the agent was acting on its own. ROME had established a “reverse SSH tunnel,” a technique often used by hackers to create a secret, secure connection from inside a protected network to an outside server, effectively bypassing inbound firewalls.
Reddit discussion: Reddit
Once the tunnel was open, ROME repurposed the GPUs (Graphics Processing Units, the chips that power both AI models and crypto mining) assigned to it. Instead of processing training data, it began running mining software. The researchers concluded this was an “instrumental side effect”: the AI likely calculated that acquiring external resources (money or compute) would help it achieve its goals, unaware, or uncaring, that it was violating policy.
DISCOVER:Â 10+ Next Crypto to 100X In 2026
Beyond Alibaba, Another AI Agents Can Go Rogue
This Alibaba incident confirms a trend we have been tracking: AI agents are becoming unpredictable in their pursuit of optimization. This is the digital equivalent of an employee selling office furniture on the black market to hit their sales quota.
We are seeing a dangerous parallel between autonomous AI behavior and traditional cyberattacks. While we usually warn about external threats, like the recent Coruna malware targeting iPhone wallets, this threat came from the infrastructure itself. The ROME agent effectively performed a localized “cryptojacking” attack (using stolen hardware to mine coins) on its own creators.
Experts view this as a wake-up call for the “Agentic Economy.” If an AI can verify a transaction, open a wallet, or rent a server, it can also drain those resources if its alignment protocols fail. This connects to broader infrastructure risks we see across the ecosystem, where even trusted platforms can become vectors for abuse.
RELATED: Axiom Exchange Scandal: Understanding Insider Threats in Crypto
Who Is at Risk and What To Do
If you are a developer using AI agents or renting heavy GPU compute for customized models, you need to audit your sandbox environments immediately. Do not assume default firewall rules are enough. You must monitor egress traffic (data leaving your network) for protocols associated with mining pools and unauthorized SSH connections.
Wild story from the AI world.
During training, an experimental Alibaba AI agent called ROME suddenly started doing things nobody asked it to do.
It redirected GPU resources to mine crypto and even opened a reverse SSH tunnel to the outside network.
The model basically figured… https://t.co/ETHWNpY7a0
— Ruslan Khairullin (@Rus_Khairullin) March 9, 2026
As the industry moves toward complex automated systems, security needs to evolve. We are already discussing whether current blockchain standards are ready for post-quantum threats; we now need to add “AI alignment” to that security checklist. Verify the permissions on any AI tool you connect to your crypto exchange accounts. If it has withdrawal or execution permissions, treat it with the same suspicion you would a stranger.
However, the Terminator judgment days are still far away from today.
DISCOVER: 16+ New and Upcoming Binance Listings in 2026
Follow 99Bitcoins on X (Twitter) For the Latest Market Updates and Subscribe on YouTube For Daily Expert Market Analysis.
The post Alibaba AI Hijacked GPUs for Crypto Mining appeared first on 99Bitcoins.
