VAST Data and CrowdStrike unveil AI-security partnership spanning infrastructure to deployment
The AMW Read
A partnership bolts security capability onto an existing AI-infrastructure player's platform, an incremental player-map update with sub-segment relevance and no funding or structural signal.
VAST Data and CrowdStrike unveil AI-security partnership spanning infrastructure to deployment
VAST Data, the AI Operating System company led by founder and CEO Renen Hallak, and CrowdStrike (NASDAQ: CRWD) announced on September 1, 2026 a partnership applying enterprise-grade cybersecurity across the infrastructure, data, and AI workloads that underpin production AI. VAST's AI Operating System now natively supports the CrowdStrike Falcon sensor, certified and available immediately. Two further integrations remain in private preview: VAST's data-access telemetry feeding into CrowdStrike's Falcon Next-Gen SIEM for threat correlation, and CrowdStrike's new AI Detection and Response product, Falcon Guardian, embedding into VAST InsightEngine — the pipeline that ingests and enriches unstructured data for AI knowledge bases — to flag PII exposure, prompt injection, and jailbreak attempts before they reach downstream models and agents. The integration is being demonstrated on NVIDIA-based AI factories, using NVIDIA's AI Data Platform reference design as the connective layer.
The partnership targets a gap opening as enterprises push AI into production: sensitive data now moves across infrastructure, users, models, and agents in ways traditional perimeter tools weren't built to observe, and data-infrastructure vendors are becoming a natural enforcement point. By wiring detection into the ingestion and enrichment pipeline rather than the model or application layer, VAST and CrowdStrike are betting that AI-specific threats are best caught where data is prepared, not where it's consumed. Per the AI Market Watch index, VAST Data has raised roughly $620 million across all rounds since its 2016 founding, though index coverage spans roughly 5,000 companies and is not a full census.
For enterprises building internal AI knowledge bases or agent fleets, this signals that data-pipeline vendors are becoming a legitimate procurement line for AI security spend, distinct from storage or retrieval performance. Builders evaluating a RAG or agent pipeline should treat ingestion-layer threat detection as separate from model-layer guardrails — and note the more consequential pieces here, SIEM correlation and Falcon Guardian, are still private preview, not general availability.
#VASTData #CrowdStrike #AISecurity #AIInfrastructure #DataPipeline #EnterpriseAI
