Chainalysis hit by AI-driven intrusion as agentic attacks spread

Key points

• Chainalysis, a blockchain analytics firm used by police and crypto exchanges, says attackers used "offensive agentic techniques", meaning AI software that runs attack steps autonomously, to reach internal systems earlier in October 2026.
• Customer names, email addresses, user IDs and role labels were accessed; the company says investigation files, passwords and session tokens were not.
• The breach lands days after Microsoft publicly tied the ConnectWise supply chain compromise to Storm-0940, a China-state aligned group using an agentic AI pen-testing framework called ARTEX.
• Chainalysis has not confirmed which framework or group was involved, and says the incident is unrelated to the recent $6.3 billion Binance theft.
• Chainalysis is using large language models to support a broader security review of its platform alongside external cybersecurity experts.

Chainalysis disclosed the intrusion on Friday evening, around 5:14 PM ET, saying an outside attacker reached internal tools and some customer records. Its software lets law enforcement, exchanges and financial institutions trace cryptocurrency payments linked to fraud and money laundering, which makes its customer list sensitive in its own right.

The stolen data is contact-level: full names, work email addresses, user IDs, and each person's role inside their organisation on the platform. No case files, login passwords or active session tokens, the digital tickets that keep a user logged in, appear to have been taken. Customers are being told to enable two-factor authentication and watch for unexpected password reset prompts or anyone calling claiming to be from Chainalysis and offering to recover lost funds.

What are "offensive agentic techniques"?

In plain terms: AI software that carries out an attack on its own. Instead of a human typing commands, an "agent" is handed a goal and picks the steps itself, scanning for weak spots, exploiting them, moving to the next machine. The appeal is speed. An agent doesn't sleep, doesn't take weekends, and can pursue several break-in paths at once.

Microsoft puts numbers on that. When we covered Storm-0940's ARTEX-driven ConnectWise campaign on 21 October, we noted the gap between first contact and privileged cloud access was under three hours in several cases. That's the tempo defenders are now being asked to match.

Is this connected to the ConnectWise and Binance incidents?

Not directly, based on public statements. Chainalysis told BleepingComputer the intrusion is unrelated to the Binance theft, which Binance has blamed on North Korean operators who compromised a DevOps engineer's laptop.

The ConnectWise angle is harder to dismiss. Microsoft attributes that supply chain compromise, where a multi-tenant application used by many customers was tampered with, to Storm-0940 using ARTEX. Chainalysis has named neither the framework nor a group. Whether its attackers used ARTEX, a different off-the-shelf agent, or something custom-built isn't known.

The attribution picture is thin. Chainalysis has shared no indicators, no CVE, no dwell time. Treat any link to Storm-0940 as low confidence until the firm or a third-party vendor publishes technical detail. What is firm is the pattern: two disclosed incidents in a single week, both describing AI-run attack chains. ARTEX has appeared in two of our stories in the past 30 days; that cadence reflects a broader shift, not isolated experiments.

Should you worry if you're a Chainalysis customer?

Chainalysis is still working out which records were touched for each customer and will send detailed notices once that analysis is complete.