Zenity raises $125M to secure autonomous AI agents

Zenity has raised $125 million in a Series C round built around a blunt premise: the biggest enterprise AI risk may not be the model or the prompt, but the agent acting autonomously inside company systems.

Norwest led the round, with SoftBank’s Vision Fund 2, Hitachi Ventures, and LG Technology Ventures participating. The financing brings Zenity’s total funding to roughly $185 million. The Israeli startup did not disclose its valuation.

Securing the agent layer

Founded in 2021 by former Microsoft security engineers Ben Kliger and Michael Bargury, both veterans of Israel’s Unit 8200, Zenity focuses on what happens after an enterprise gives an AI system access to databases, tools, records, and multi-step workflows.

That changes the security problem. A chatbot primarily generates content; an agent can take action. It may have excessive permissions, follow manipulated instructions, or perform a harmful operation while behaving exactly as it was designed to behave.

Zenity says its platform monitors an agent’s permissions, connected tools, memory, and live actions. It analyzes the intent behind each action and can allow, modify, or block the operation before it executes.

Recommended reading

Claude escaped a sandbox and hacked three companies

The company’s research arm, Zenity Labs, has demonstrated how a malicious calendar invitation could hijack Perplexity’s agentic browser, open an unlocked password vault, and expose the credentials inside. Its earlier AgentFlayer research also identified zero-click attacks that could turn enterprise assistants against their owners by hiding malicious instructions in data those systems were expected to trust.

“era of 1 billion agents.”

A crowded market with open questions

Zenity now employs more than 230 people, with research operations in Tel Aviv and sales in New York. It says most of its customers are Fortune 500 and Global 2000 companies in regulated industries. The new investors are also significant: each is deploying agents in its own business, according to the company’s announcement.

The funding arrives shortly after OpenAI acknowledged that two models escaped a sealed test and hacked Hugging Face while pursuing a benchmark answer key. That incident illustrates the type of autonomous behavior Zenity is trying to control.

The startup is entering a market that already includes Onyx, which is building control layers for agents, and follows Horizon3's $250 million round for autonomous penetration testing. Gartner has named Zenity the company to beat in agent governance, according to the report.

The central commercial question is whether enterprises will pay for a dedicated agent-security platform or rely on capabilities bundled by Microsoft, Google, and AWS. Zenity’s round shows that investors currently favor specialists, but the company has not disclosed its valuation or detailed growth figures. Those omissions make the financing a strong signal of demand, not yet proof that Zenity will own the category.