The data layer is critical for AI to be fully exploited by enterprises. To protect this infrastructure, VAST Data argues that built-in security is required. Thanks to a collaboration with CrowdStrike, achieving the desired level of protection has become one step easier.
Through the collaboration, VAST Data provides threat detection, automated response, and workflow security for AI and analytics environments within its own VAST AI OS. For VAST, the partnership represents an extension of its security-first approach to enterprise AI. The company is introducing a new generation of fully observable and controllable AI infrastructure. It features unified policy management, agent governance, encryption, and regulatory reporting.
VAST Data founder and CEO Renen Hallak argues that AI systems are only as reliable as the data they rely on. To keep that data free from manipulation and leaks, built-in security is needed. The integration with CrowdStrike brings the usual features of that security player into the context of VAST Data’s platform.
Real-time detection at the data layer
The integration works across several layers. VAST continuously monitors data access, administrative activity, and workload behavior. These signals are correlated with CrowdStrike’s endpoint, identity, and threat intelligence data for real-time detection. When a risk is identified, CrowdStrike can use secure API actions within VAST to quarantine the data, restrict access to it, isolate workloads, and preserve forensic records.
The two parties also bring security directly to companies’ workflows. Files and objects can be scanned, analyzed, and approved before they enter enterprise pipelines. This prevents compromised data from reaching AI models and analytics systems. It reduces the risk of model poisoning, data leaks, and unauthorized access. The consequences of these problems can be subtle or even go unnoticed, but that doesn’t make them any less dangerous. Consider the influence of AI models in very specific circumstances, where AI tooling with too much access to data can assist in a leak.
Open design
Although the collaboration is based on deep integrations, both parties emphasize the openness of the design. To drive adoption, VAST Data and CrowdStrike offer standardized APIs, open connectors, and reference architectures to enable customers and partners to quickly use the combined tooling. This is also necessary to integrate IT security for the data layer with the security of all other components of an IT environment. This allows IAM solutions, alternative XDR tools, and cloud security applications, among others, to use the security data generated around the data layer itself.
Read also: Who will develop the OS for AI? VAST Data is going for it