Infisical raises $40M to secure AI agent credentials
The security startup is building infrastructure to manage secrets and permissions for non-deterministic AI agents that scale exponentially.
The short version
- Infisical raised $40 million at a $600 million valuation in a Series B round led by Elad Gil.
- The platform focuses on 'non-deterministic' behavior, where AI agents might perform unintended actions.
- The company uses an open-source model to allow for transparent code inspection and build developer trust.
- Infisical provides a single-pane platform for secrets, certificates, and access management across cloud and on-premises environments.
Why it matters
As companies move from human-led tasks to AI agents that can perform actions autonomously, traditional security models often fail. This funding marks a shift toward specialized infrastructure that treats AI agents as distinct entities requiring their own isolated permissions.
Funding for agent security
Infisical, a security infrastructure startup based in San Francisco, recently secured $40 million in Series B funding. The investment brings the company's total funding to $58.93 million and sets its valuation at $600 million. This capital will be used to advance security for AI agent operations, specifically focusing on credential isolation and approval workflows.
The funding round was led by Elad Gil, who also led the company's previous $16 million Series A round in June 2025. Infisical was founded in 2022 and is currently led by CEO Vlad Matsiiako, a Cornell University graduate with a background in operations research and information engineering. The company currently employs 43 people.
The primary goal of this new capital is to address the unique challenges posed by AI agents. Matsiiako noted that many security primitives that worked for human developers are no longer scaling for the high volume of traffic and operations performed by autonomous agents. The company plans to offer a suite of tools including secrets management, certificate management, and access management.
The problem of non-determinism
A core challenge Infisical identifies is the non-deterministic nature of AI agents. Unlike humans or traditional machines, which follow predictable paths, agents can sometimes 'wander' and perform unintended actions when given a task. This unpredictability makes it difficult to manage permissions using standard security protocols.
Because agents can scale exponentially, they create a massive increase in secrets and infrastructure touchpoints. If an agent has broad access, a single unintended action could lead to significant security breaches. Infisical aims to solve this by providing isolated or ephemeral credentials, which are temporary keys that expire after a task is completed.
Matsiiako explained that the difference lies in how agents behave compared to previous machine operations. He noted that agents can perform all types of actions and might do something unintended. This behavior requires a more granular level of control than what is currently available in many legacy tools.
Open source and trust
Infisical utilizes an open-source model to build trust with its customers, particularly large enterprises and regulated organizations. By making the code open, the company allows users to inspect the security infrastructure themselves to ensure it is robust. This transparency helps reduce the vendor trust risk that often accompanies closed-source security products.
The company differentiates itself from competitors by offering a natively integrated single-pane platform rather than a collection of stitched-together acquisitions. This approach is designed to reduce overhead and accelerate setup for organizations. It allows for the enforcement of uniform security policies across both cloud and on-premises environments.
The goal is to create a frictionless adoption process for both early-stage startups and large banks. By providing a transparent product, Infisical hopes to make it easier for security teams to approve the tools that developers actually want to use. This transparency is intended to facilitate faster adoption across diverse technical stacks.
Aligning dev and security teams
Infisical focuses on developer experience as a primary driver for security adoption. The company believes that security tools only work if developers and agents actually use them correctly. If the tools are too difficult to use, developers will often find workarounds that create security holes within the organization.
To prevent this, Infisical builds in features like built-in integrations, APIs, and UI-based approval workflows. These features are designed to provide efficiency gains for developers while maintaining the oversight required by security teams. Matsiiako noted that the platform aims to put security and developer teams on the same team by making the right way to do things the easiest way.
The platform supports a wide range of integrations, including cloud providers, CI/CD tools, Databricks, and sandbox providers. By centralizing credentials across these heterogeneous stacks, Infisical helps manage sensitive resources in complex environments. The company emphasizes that everything is designed to be within the same dashboard to simplify the management of the entire development-to-production life cycle.
What to watch
- How other security firms respond to the shift toward agent-specific credential isolation.
- Whether open-source security models become the standard for regulated industries like banking.
Try it
Check if your current secrets management tool supports 'ephemeral' or 'just-in-time' credentials for automated processes.
Sources and links
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by BankInfoSecurity · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
