Cyera, a data security platform, is acquiring Oasis Security for $1 billion in an all-stock deal. The acquisition represents Cyera's third major purchase in 2024 and signals growing concern about protecting AI agents in enterprise environments.
Oasis Security specializes in securing AI agents, autonomous systems that perform tasks without human intervention. As companies deploy more AI agents across operations, the attack surface expands. Oasis provides visibility and control over these systems, detecting unauthorized behavior and preventing data exposure.
The combination addresses a real gap. Data security tools designed for traditional workloads don't handle AI agents effectively. Agents operate autonomously, access sensitive data, and interact with APIs unpredictably. Oasis' platform monitors agent behavior, identifies anomalies, and enforces security policies specific to autonomous systems.
Cyera's CEO Lior Div stated the acquisition strengthens the company's ability to protect data across the entire AI lifecycle. The platform will integrate Oasis' agent security capabilities with Cyera's existing data governance tools, offering enterprises unified visibility from data storage through AI model inference.
This deal reflects the maturation of AI security concerns. Early AI adoption focused on model safety and hallucinations. Now enterprises confront practical deployment challenges: controlling where agents access data, preventing prompt injection attacks, and auditing agent decisions for compliance.
Cyera has been aggressive on acquisitions. Earlier in 2024, the company bought Wiz's data security division and another unnamed startup. These moves suggest Cyera is positioning itself as a comprehensive data protection platform for the AI era, covering everything from data classification to AI-specific threats.
The $1 billion valuation for Oasis reflects investor conviction that AI agent security represents a standalone market opportunity. Startups across security have raised significant funding on the premise that existing tools fail against AI-specific risks.