The extortion group FulcrumSec has published roughly 1.3 terabytes of confidential files allegedly stolen from Danish pharmaceutical giant Novo Nordisk after the company refused a $25 million ransom demand. Novo Nordisk publicly confirmed a security incident on June 11, 2026, acknowledging unauthorized access to a restricted set of internal technology systems and the exposure of personal information. FulcrumSec claims the haul spans more than 700,000 files, including source code, clinical research, and proprietary AI system data.
What Happened
According to FulcrumSec, the intrusion began in March 2026 when the group obtained a valid GitHub authentication token. That single credential opened the door to internal software repositories, which the attackers used to harvest further logins and expand their foothold across Novo Nordisk's environment. The group says it operated undetected for more than two months, quietly staging and exfiltrating data before making contact.
FulcrumSec reached out to undisclosed Novo Nordisk executives with a $25 million payment demand. The company responded on June 3, roughly 48 hours after initial contact, authenticating its identity through a Proton Mail account before ultimately rejecting the terms. Novo Nordisk went public with the breach on June 11. With the ransom refused, FulcrumSec followed through on its threat and began releasing and marketing the stolen material.
What Was Taken
FulcrumSec claims to have exfiltrated approximately 1.3 terabytes of data comprising more than 700,000 individual files. The group describes the contents as including:
- Source code from internal software repositories
- Confidential pharmaceutical research and clinical study documentation
- Proprietary AI system files
- Personal information, which Novo Nordisk confirmed was exposed
Notably, FulcrumSec states it intends to pursue targeted private sales of select portions of the data while pledging to withhold patient information, employee records, and production facility data. That claimed restraint is unverified and self-serving, and it should not be treated as a guarantee.
Why It Matters
For a pharmaceutical company, the loss of clinical study documentation, research pipelines, and AI system files represents a direct threat to competitive advantage and years of R&D investment. Stolen source code can seed follow-on attacks by revealing internal architecture, hardcoded secrets, and undocumented endpoints. The confirmed exposure of personal information carries regulatory weight under GDPR and similar frameworks, with potential notification obligations and penalties.
The episode also reinforces a broader pattern: extortion groups increasingly bypass encryption entirely, opting for pure data theft and public leak pressure. When a victim refuses to pay, the fallout is a slow-burn release and quiet private sales rather than a single dump, extending the damage window for defenders, partners, and regulators alike.
The Attack Technique
The reported root cause is a leaked GitHub access token discovered by the attackers in March. Exposed source-control credentials, whether committed to a repository, embedded in a build artifact, or leaked through a third party, remain one of the most common and damaging initial-access vectors. Once inside the repositories, FulcrumSec claims it mined additional credentials to move laterally and deepen access, a classic credential-chaining progression that turned one token into broad network reach and more than two months of undetected persistence.
What Organizations Should Do
- Audit and rotate all source-control tokens: enforce short-lived, scoped access tokens for GitHub and equivalent platforms, and revoke long-lived personal access tokens.
- Deploy secret scanning across repositories, commit history, CI/CD pipelines, and build artifacts to catch exposed credentials before attackers do.
- Enforce phishing-resistant MFA on all developer and repository accounts, and disable legacy authentication paths that bypass it.
- Monitor for anomalous repository access, mass clone or download activity, and unusual data egress volumes that indicate staging and exfiltration.
- Apply least privilege to service accounts and tokens, segmenting repository access so a single compromised credential cannot reach the entire codebase.
- Rehearse a no-pay extortion playbook in advance, including legal, regulatory notification, and communications workflows, so a refusal decision does not become an improvised scramble.