Rein Security Secures $25M Series A to Expand Runtime Defense for AI Agents
Rein Security raised $25 million in Series A funding to expand its runtime sidecar technology, which monitors and protects autonomous AI agents across enterprise environments.
Rein Security has raised $25 million in a Series A financing round to scale its runtime protection platform built specifically for enterprise artificial intelligence agents. The round was co-led by Glilot Capital and Sienna Venture Capital, with backing from Corner Ventures, Atlacle, and RNP Capital Advisors, increasing the company's total funding to $35 million.
Founded in 2024 and co-headquartered in New York City and Tel Aviv, Rein Security focuses on addressing the emerging vulnerabilities introduced when autonomous software systems interact directly with internal corporate workflows, data stores, and third-party tools. As organizations deploy agents capable of executing multi-step tasks across disparate applications, standard security defenses often fail to inspect or intercept downstream actions triggered by machine-to-machine interactions.
Securing Agent Execution at Runtime
Unlike traditional security tools that rely on network gateways or API proxies, Rein deploys patented sidecar technology that runs directly where an AI agent executes actions. This runtime setup allows security teams to inspect every line of code executed by an agent and track every business resource touched, without having to route internal or customer data through an external proxy.
According to Gartner, enterprise spending on securing AI is expected to reach nearly $4.8 billion in 2027—a 68.7% rise compared to 2026—and expand further to roughly $7.7 billion by 2028. The surging demand highlights how enterprise teams are grappling with two simultaneous pressures: protecting the proprietary agents they deploy internally and shielding their systems from automated, adversarial AI attacks operating at machine speed.
Rein's vulnerability research group, Agent Breakers, illustrated these practical risks at Black Hat USA 2026, demonstrating how researchers compromised the autonomous shopping agent of a top-five retailer in the United States. In another production environment cited by the company, an enterprise onboarding agent ingested an ordinary-looking PDF containing a hidden prompt injection designed to push the agent outside its authorized scope. Rein’s runtime guardrails identified and intercepted the malicious instruction before unintended actions could be carried out.
Enterprise Adoption and Market Traction
Since its launch in January 2026, Rein Security reports an eightfold increase in revenue and a fivefold expansion in its customer base. The platform currently monitors thousands of AI agents executing millions of discrete actions across industries including SaaS, financial services, healthcare, retail, data, and energy.
Its enterprise roster includes companies such as Flex, Swimlane, Dun & Bradstreet, and insurance platform Lemonade. Rein's guardrails protect Lemonade's customer-facing services supporting more than three million active users, as well as agentic features at Dun & Bradstreet that serve over 240,000 customers.
"Our customers rely on Dun & Bradstreet for trusted data, and that trust has to extend to every AI agent we put into production," said Jay DePaul, chief cybersecurity and technology risk officer at Dun & Bradstreet. "Rein helps us see and control what our agents do at scale, so we can keep innovating with agentic AI while protecting the customers who depend on it."
Rein plans to allocate the newly raised capital toward accelerating core product development, expanding agentic vulnerability research, and hiring globally.
What it means for developers
For engineering teams building autonomous pipelines, Rein's funding highlights a fundamental transition in how AI systems must be secured. Early generative AI deployments primarily required input filtering and basic prompt validation. However, as developers connect models to databases, terminal environments, and enterprise APIs, security shifts from managing prompt posture to governing active code execution.
Because agents frequently act on context received from secondary sources—such as customer uploads, emails, or output from other models—engineers must account for indirect prompt injections that instruct agents to perform unauthorized actions downstream. Sidecar and runtime monitoring architectures offer a way to trace every action back to the specific execution step without adding significant latency or routing confidential data to outside proxies.
When prototyping agentic workflows across different foundation models, developers can try top AI models cheaply through one API at https://apixoai.online to evaluate how various architectures handle complex reasoning and tool usage. Coupling diverse model evaluations with robust execution guardrails will be critical as agent autonomy moves deeper into production systems.
Source: Rein Security Raises $25M to Secure the AI Agents Enterprises Build and Stop the Ones That Attack Them — AI Insider. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
