Pluto Security Emerges From Stealth With $42M to Tackle AI’s New Enterprise Risk

Pluto Security has emerged from stealth with $42 million in funding as enterprises scramble to manage a fast-growing security problem: employees are adopting AI tools and building AI-powered workflows faster than corporate defenses can track them. The startup says the new money will help it expand its product, deepen research, and scale in the U.S. as it tries to secure what it calls the shift from human behavior to agent behavior.

The timing is notable. Pluto’s co-founder and CEO, Shahar Bahat, says new AI tools can appear inside a company environment within hours of release, creating a widening attack surface that traditional security platforms were not designed to monitor. The startup argues that the biggest risk is no longer just what employees do with sanctioned apps, but what those employees create, connect, and expose when they use AI to build software and automate work.

Why Pluto thinks enterprise security needs a reset

Pluto Security’s pitch is built on a simple premise: widespread access to AI tools has changed the nature of corporate security, and defenders need visibility into AI usage in real time rather than after the fact. Bahat says many companies are now encountering risky or malicious behavior inside AI-generated tools, extensions, and agent workflows, often before security teams even know the tools exist.

According to the company, about 60% of the risks it detects across customers come from tools built by non-developers using so-called vibe coding. Finance teams are the most frequent source of those issues, followed by marketing and legal groups, while professional developers rank fourth. In one example, the company says it found a finance employee at a Fortune 500 firm had created an app to manage customer information and churn, but had left it exposed to the public internet.

The message to security teams is clear: AI adoption is no longer confined to engineering departments, and the resulting shadow IT is becoming shadow AI.

How Pluto Security works

Pluto’s software is designed to provide visibility, risk analysis, and control over how employees use AI tools and how they assemble AI-driven systems. The company says its platform monitors the tools people use, the extensions and skills those tools rely on, the Model Context Protocol servers involved, and the agents or applications employees create.

Rather than just cataloging activity, Pluto says it can also intervene while work is underway. If an AI agent attempts to gain permissions beyond what a user intended, the system can halt the action or inject a message into the session prompting the agent to ask for confirmation.

Bahat described the company’s approach as a shift from monitoring user behavior to monitoring agent behavior, arguing that defenders now need to understand what autonomous systems are doing as they operate.

To do this, Pluto integrates with endpoint security and device management systems. It can also place its own sensors on endpoints, either through those management platforms or directly on devices. The company says it collects information such as prompts, commands, active processes, permissions, and earlier user actions, then uses lightweight models on the endpoint to help infer what an AI agent is trying to do.

What makes Pluto’s approach different?

What sets Pluto apart, according to the company, is that it does not rely only on after-the-fact logs or broad policy enforcement. Instead, it attempts to make decisions in the middle of an AI session, when an agent is actively requesting access, taking actions, or chaining together tools in ways that might be risky.

This is a meaningful distinction for security teams that have long struggled with the balance between locking systems down and letting workers use new productivity tools. Pluto says its system aims to enforce policy without stopping employees from experimenting with or deploying AI.

• Monitors AI tools, connectors, extensions, and agent activity
• Identifies risky components and suspicious behavior
• Intervenes in real time when access requests look unsafe
• Works through endpoint and device management systems
• Uses local models to interpret what agents are trying to do

How did Pluto reach $42 million so quickly?

Pluto reached this funding milestone rapidly because it entered a market where enterprise buyers are increasingly anxious about AI-related exposure. Investors have been willing to back startups that promise to secure the new layers of software now being created by employees across the business, not just by central IT or engineering groups.

The company’s financing includes an $8.5 million seed round led by TLV Partners, with participation from Armis founders. Pluto later raised a $34 million Series A led by Jump Capital, with additional backing from Skinos, the venture firm founded by Check Point co-founder Shlomo Kramer, as well as Ansa Capital and other angel investors.

That capital structure gives Pluto both early validation and room to move quickly in a highly competitive market where security startups are racing to define the category around AI governance, AI visibility, and AI agent defense.

Pluto’s growth so far

Although the company only started selling in April, Bahat says Pluto has already reached more than $2 million in annual recurring revenue and signed over 30 enterprise customers. Those customers include Oscar Health, Workato, Sixt, and AlphaSense.

For a startup that only recently began commercial sales, that is a strong signal that buyers are already spending on tools they believe can help them understand AI risk before it becomes an incident. Pluto says its traction is not just coming from outbound sales. Bahat says roughly 70% of the company’s pipeline now comes from inbound interest, fueled in part by its free public-facing platform.

What is Plutonium, and why does it matter?

Plutonium is Pluto’s free community platform for checking the risks associated with publicly available AI skills, connectors, and add-ons. The company also offers an MCP scanner and a catalog of verified components, giving security professionals and builders a way to inspect the ecosystem around AI tools before rolling them out internally.

Bahat says thousands of security professionals use the platform, and that the community around it has become a major source of feedback and customer leads. In effect, Plutonium acts as both a research tool and a funnel for the commercial product.

This dual strategy matters because many enterprise security startups struggle to prove relevance before procurement cycles drag on. By giving away a useful tool up front, Pluto is trying to establish trust, collect data about what practitioners are worried about, and make itself part of the conversation before a company is ready to buy.

Why security teams are worried about AI agents

Security leaders are increasingly concerned that AI agents can make changes, call tools, and access data in ways that are difficult to predict, especially when employees assemble them from third-party extensions, connectors, or open components. That is what makes the current wave of AI adoption different from earlier software sprawl: the software is not just deployed, it is also composed on the fly by workers outside the IT department.

Pluto says its customers are finding malicious behavior in almost every proof of concept it runs. Bahat says that during evaluations the company repeatedly encounters dangerous extensions or skills, and in some cases even active attacks while the proof of concept is underway.

Bahat said that in every proof of concept the company has entered, it has uncovered some form of malicious activity, ranging from suspicious extension packages to live attacks occurring during testing.

That claim, if representative of the market, underscores how early many companies still are in understanding the security implications of AI tooling. It also helps explain why startups like Pluto are attracting funding at pace: buyers are under pressure to get ahead of risk before employees normalize unsafe AI usage.

Who is buying these tools?

Pluto’s initial customer base suggests that demand is coming from a broad mix of industries rather than only from highly technical cloud-first companies. The company says its customers include Oscar Health, Workato, Sixt, and AlphaSense, illustrating interest from health care, software, mobility, and data-driven business services.

That broad adoption pattern is important because it shows the problem is not limited to one sector. Any company that lets employees experiment with AI copilots, no-code builders, browser extensions, or agent frameworks may need new controls.

As Bahat sees it, the intensity of risk depends partly on how open an organization is to AI and how its security leadership approaches adoption. But the basic exposure is spreading regardless of sector.

Funding, traction and market context

Pluto’s fundraising comes at a moment when the security market is adjusting to the rise of generative AI and autonomous agents. A growing number of startups are promising to help enterprises govern, monitor, and control employee use of AI tools, but many incumbent security systems were built for a different era of risk.

The company believes that means traditional monitoring alone is insufficient. The new stack, in Pluto’s view, must understand what is happening inside AI sessions as they unfold, not just whether a user has installed a tool or signed up for a sanctioned service.

What Pluto plans to do next

Pluto says it will use the new capital to improve its product, increase security research, and grow its U.S. footprint. The company currently employs about 40 people split between the U.S. and Israel.

Its challenge now is to translate a sharp market thesis into a durable category position. That means proving that AI agents can be monitored and constrained without crippling the productivity gains that made them attractive in the first place. It also means convincing security buyers that the company can stay ahead of a landscape where new tools, connectors, and model-driven workflows appear constantly.

For now, Pluto is betting that enterprises will want more than approval lists and usage logs. They will want a security layer that watches how AI behaves, blocks dangerous actions, and helps employees use new tools without opening the door to data leaks or unauthorized access.

Why this startup matters now

Pluto is arriving at a moment when the gap between AI adoption and corporate control is becoming impossible for many businesses to ignore. Employees are increasingly able to build applications, automate work, and connect systems with little technical oversight, and that creates a new class of security risk that conventional tools may not catch.

Whether Pluto becomes a category-defining company or simply one of several winners in an emerging market will depend on how well it can prove its technology works at scale. But its early traction, its strong inbound demand, and its focus on real-time intervention suggest that the market for AI security is only just beginning to take shape.

For enterprise buyers, the message is even simpler: AI has become a security issue, not just a productivity story.