Pluto Security's $42M Play for Enterprise AI Agent Security

Pluto Security exited stealth on October 13, 2026, with $42 million in total funding and a blunt diagnosis of the enterprise: the people adopting AI agents fastest are the same people least equipped to secure them. The startup is betting that enterprise AI agent security is the next major category in cybersecurity, and its early numbers suggest buyers agree.

The round is split between an $8.5 million seed led by TLV Partners, with participation from the founders of cybersecurity firm Armis, and a $34 million Series A led by Jump Capital. Skinos, Ansa Capital and various angel investors also participated, according to TechCrunch, which first reported the news.

The vibe-coding security gap

The core problem Pluto is chasing starts with who builds AI software now. According to Security Boulevard, tools like Claude Code, Codex, Cowork, Lovable and Base44 let employees far outside engineering teams build applications and automate processes, without following traditional development workflows. That democratization is the feature. The security hole is the side effect.

Pluto's chief executive and co-founder Shahar Bahat told TechCrunch that a newly announced AI tool can show up inside a company's environment within two hours of its release. Speed of adoption is one problem. Authorship is the other: the company says about 60 percent of the risks it detects originate from tools vibe-coded by non-developers, with finance departments the biggest source, followed by marketing and legal. Developers rank fourth.

That ordering inverts the classic application-security model, which assumes engineering teams are the ones shipping risky code, and it defines the shape of the enterprise AI agent security problem. Pluto found something else. In one case reported by TechCrunch, an app built by a finance employee at a Fortune 500 company to manage customer information and churn was left exposed to the public internet. The app worked. Its permissions did not.

What Pluto actually sells

Pluto's platform gives enterprises visibility, contextual risk assessment and real-time control across the AI tools employees use, the supply chain surrounding those tools, and the applications, agents and automations employees create, according to Security Boulevard.

The company frames the threat as an AI supply chain problem. Its scanners map skills, extensions, packages, model context protocol servers, models and other components that Pluto says number in the hundreds across its customer base, feeding thousands of employee-built applications. In April, Pluto released a free scanner called Plutonium that checks AI skills, connectors and add-ons for risks, according to AI Weekly.

The pitch rests on a striking claim about what happens when that scanner runs inside a prospective customer. Bahat told TechCrunch that in every proof of concept the company has run, it found something malicious inside the customer's AI deployments, including active live attacks. The product also supports real-time agent monitoring and mid-session intervention, meaning it can act on dangerous agent activity while it is happening rather than auditing it afterward. That enforcement capability is what separates an enterprise AI agent security platform from a conventional audit trail.

The numbers behind the bet

Pluto began selling only in April and has already crossed $2 million in annual recurring revenue, according to TechCrunch. The company says it has signed more than 30 enterprise customers, including Oscar Health, Workato, Sixt and AlphaSense, and secures more than 1 million endpoints globally. Roughly 70 percent of its sales pipeline comes from inbound interest, which is unusual for an early-stage vendor in enterprise AI agent security and suggests the pain is doing its own prospecting.

The company has 40 employees across the United States and Israel. It plans to use the new capital to accelerate product development, expand its security research efforts and grow its business presence within the United States, according to the announcement. More details are available on the company's website, pluto.security.

Why this matters for the agent economy

Pluto's emergence fits a pattern worth watching for anyone building on agents: the infrastructure around the agent economy is maturing in layers, and the trust layers are getting funded right after the capability layers. Enterprise AI agent security is moving from a nice-to-have to an expected control. Harbor recently raised $10 million to offer insurance for AI agent failures, treating agent mistakes as an insurable risk class. Anthropic's new distillation safeguards cut model-extraction attacks by 90 percent, hardening the model layer itself.

Pluto sits between those two: it hardens the deployment layer, where agents meet real enterprise data and real employee improvisation. The challenge for security teams, as Security Boulevard puts it, is to understand the activity surrounding these tools without preventing employees from using them productively.

That framing explains why Pluto avoids the language of blocking and locking down. The company's own data says the riskiest builders are finance teams, not engineers, so the buyer is not the application-security team policing a code repository. It is the security team trying to govern an environment where any employee can ship software on a Tuesday afternoon. As reported by TechCrunch, investors are wagering that existing security software was not built for that world. Pluto Security just raised $42 million to prove that enterprise AI agent security is its own discipline.