AI evaluation startup Braintrust disclosed a security breach affecting its Amazon cloud environment. The company is instructing all customers to immediately rotate their API keys.
Braintrust, which provides infrastructure for engineers building AI software, notified its customer base of unauthorized access to one of its cloud systems hosted on Amazon Web Services.
The startup has not disclosed the full scope of the breach or what data may have been accessed. However, the decision to mandate API key rotation across its entire customer base suggests the attackers gained access to sensitive authentication credentials.
API keys are critical authentication tokens that allow applications and users to access cloud services and APIs. Rotating these keys—essentially replacing old credentials with new ones—is a standard security practice to prevent unauthorized access if credentials are compromised.
Braintrust's breach comes amid increased scrutiny of AI company security practices. Startups in the space handle sensitive model data and customer information, making them attractive targets for attackers seeking to steal proprietary AI systems or access credentials.
The company has not disclosed when the breach occurred, how long attackers maintained access, or whether customer data beyond API keys was compromised. Details about the investigation and remediation efforts remain limited.
Customers using Braintrust's platform have been advised to prioritize API key rotation. The company has likely provided instructions for generating new credentials and updating their systems accordingly.
This incident underscores the security challenges facing emerging AI infrastructure companies. As these startups become more central to AI development workflows, they also become higher-value targets for cyber attacks.
Braintrust has not released a detailed incident report or timeline. Further details about the breach's extent and impact may emerge as the company completes its investigation.
U.S. federal prosecutors have unsealed charges against three Russian nationals accused of operating a bulletproof hosting service that supported ransomware gangs responsible for over $62 million in damages worldwide.
The U.S. Cybersecurity and Infrastructure Security Agency (CISA) warned that attackers are actively exploiting three vulnerabilities in Internet-exposed on-premises SharePoint Server instances. Organizations running affected versions must patch immediately.
Tailscale disclosed a critical vulnerability in its SSH implementation that allowed attackers to gain root access through insecure argument handling. The flaw has been patched in recent versions.
A new study found that social media platforms referred over 5.7 million visits to nonconsensual deepfake pornography sites between December 2025 and March 2026, with YouTube and X accounting for the majority of traffic.