Chainalysis: credentials and tokens stolen from all of its customers
An attack using AI-assisted techniques compromised Chainalysis systems and exposed credentials, hashes, and tokens of all its users. What it means for your SMB.
4 min read

In this article
The blockchain analytics firm Chainalysis has confirmed a security incident that, given its scope, should make any IT team stop and think. According to the company's acknowledgment, the attackers obtained sensitive personal information from all of its users: names, usernames, password hashes, and active tokens. We are not talking about a handful of affected accounts, but the entire set of customers provisioned on its platform.
The detail is not minor. Among the customers that the company itself and public sources have identified are government agencies such as the FBI, U.S. ICE, and the U.S. Air Force, as well as private companies like Microsoft, Coinbase, and Barclays. That a firm that sells trust and traceability has suffered a breach of this magnitude is, at the very least, uncomfortable.
What exactly happened
Chainalysis reported last Friday that it had been the target of an incident in October, and described it as similar to that of other platforms attacked that same month. The external actor used offensive techniques assisted by AI agents to gain unauthorized access to its systems and internal tools. In other words, this is not classic phishing or an isolated exploit: the attacker automated part of the dirty work with artificial intelligence.
In the email sent to affected customers, the company states that "non-sensitive customer data, such as user identifiers, full names, email addresses, and organization role labels" was accessed. It is striking that the company classifies that package as non-sensitive. A full name, a corporate email, and a role within the organization are excellent raw material for a well-constructed social engineering attack. And the password hashes and active tokens, although not mentioned in that sentence, also traveled in the batch according to its own initial acknowledgment.
Based on the available information, it is not clear whether the data theft has exposed customers to additional threats. That uncertainty, in security, is usually as bad news as certainty.
Why this matters to an SMB
It may seem like a problem for large accounts, but the pattern is the one we see every week in small and medium-sized companies. Three concrete lessons:
- Tokens are master keys. An active API token does not expire just because. If it leaks, the attacker gets in without needing to break a password. Reviewing token expiration and rotation should be routine, not a pending task.
- Hashes are not harmless. With a hash in hand, an attacker can attempt offline cracking. If the password is weak or reused, the hash becomes a password in a matter of hours.
- AI attacks too. The use of AI agents for offensive tasks is no longer science fiction. A small team cannot compete on volume of attempts, but it can compete on hygiene: two-factor authentication, secrets management, and monitoring.
What we recommend doing starting today
When faced with a provider that suffers a breach of this kind, the reaction cannot be to wait and see what happens. We propose a simple shock plan that can be executed in an SMB:
- Integration inventory. Locate all tools that connect to external platforms via tokens or API keys. If there is no inventory, there is no control.
- Immediate rotation. Change the tokens and credentials associated with any affected provider. Do not wait for the provider to ask.
- Password review. If hashes have been exposed, force a password change and check that they are not reused on other services.
- Two-factor everywhere. No exceptions for email, VPN, administration panels, and SaaS tools.
- Access monitoring. Review login logs and anomalous behavior alerts over the coming weeks.
This incident fits the trend we already discussed in connection with the critical flaw in OpenSSH and musl with active attacks: the window between the publication of a vulnerability and its exploitation has narrowed. It also recalls what happened with the three WhatsApp flaws, one of them already used in scams: a rapid response marks the difference between a scare and a serious incident.
Trust in a provider is not inherited: it is verified. And when it fails, the speed of your response is your only real firewall.
At ForgeNEX we insist on something unglamorous but effective: treating credentials and tokens as critical assets, with inventory, rotation, and alerts. You do not need an elite cybersecurity department to do it right; you need consistency and a written procedure. If your SMB depends on external integrations and you do not know how many active tokens you have right now, that is the first problem to solve this week.
Source: Ars Technica. Analysis and adaptation: ForgeNEX.
Keep reading
- WhatsApp fixes three flaws: the Android one has already been used in scams22 November 2026
- Critical flaw in OpenSSH and musl: active attacks already underway21 November 2026
- Zero-day in VMware vCenter: patch now or assume the risk20 November 2026
- Alpine and musl flaw: passwordless SSH login under certain configurations19 November 2026
- WordPress 7.1.5 fixes 20 flaws: unauthenticated RCE and critical SQL19 November 2026