Asymptote Labs, a startup focused on building the critical runtime control and visibility infrastructure necessary for the secure deployment of enterprise AI agents, has announced the successful closure of a $5.5 million seed funding round. The investment arrives at a pivotal moment for corporate security teams, who are increasingly struggling to maintain oversight as autonomous agents gain deeper access to sensitive internal systems and proprietary data.
The oversubscribed financing round was led by Tower Research Ventures. The startup also attracted participation from a group of strategic investors, including Lytical Ventures, 1984 Ventures, and South Park Commons. This infusion of capital is earmarked to accelerate the company’s product development roadmap, facilitate the expansion of its engineering team, and bolster its go-to-market strategies as it looks to establish a foothold in the rapidly evolving landscape of AI security.
Asymptote Labs was founded by CEO Justin D’Souza and CTO Shukan Shah. The pair brings relevant experience to the challenge, having previously worked on security and defensive infrastructure at the AI-driven platform Doppel. Their firsthand experience in the field led them to identify a significant and growing gap in the market: while enterprises are rushing to adopt autonomous agents to drive efficiency, there is a profound lack of transparency and control over what these agents are actually doing once they are unleashed within a corporate environment.
Security teams are currently facing severe "blind spots" when it comes to autonomous tools. Because these agents are designed to act with varying degrees of autonomy—often interacting directly with APIs, reading sensitive files, and executing code—they present a unique risk profile. Traditional security tools, which were largely designed for static applications or human-user behavior, often fail to interpret the intent or the runtime consequences of an AI agent’s actions. D’Souza and Shah founded Asymptote to bridge this gap, providing the "eyes and ears" that security operations centers (SOCs) currently lack.
The foundation of the company’s solution is Beacon, an open-source telemetry layer designed to standardize the way agent activity is monitored. In the current fragmented AI ecosystem, there is no single standard for how agents report their actions. Developers and security teams are forced to grapple with a wide variety of agent harnesses, each of which logs data in different formats and granularities. Beacon seeks to solve this by providing a unified telemetry layer that supports over 20 different agent frameworks, including popular tools such as Cursor, Codex, and Claude Code. By normalizing this telemetry into a cohesive data model that extends the industry-standard OpenTelemetry framework, Beacon allows organizations to ingest and analyze agent activity data without having to build custom integrations for every new tool they adopt.
However, Asymptote Labs is not merely an open-source contributor; the startup has developed a commercial enterprise platform that functions on top of this open-source core. This platform is designed to provide active, real-time security for agent operations. While Beacon offers the visibility, the enterprise platform adds the necessary control mechanisms to prevent incidents before they manifest as data breaches or unauthorized system changes.
The platform functions by intercepting live agent activity, allowing security teams to enforce policy guardrails at both the prompt level and the tool execution level. By analyzing the "intent" of an agent’s request before it interacts with a database or a production environment, the system can steer agents away from unsafe behavior. If an agent attempts an action that violates corporate policy—such as attempting to access a forbidden file path or executing a suspicious shell command—the platform can block the action in real-time. This proactive approach to security is a marked shift from traditional post-mortem analysis, where security teams only discover a breach after the fact.
"AI agents are becoming a new operating layer inside every company, but the infrastructure to understand and control them hasn’t caught up," said D’Souza. "We saw this problem firsthand when building agents at Doppel. As companies deploy thousands of agents across their infrastructure, their runtime data will become one of the most valuable assets they own."
The urgency behind Asymptote Labs’ mission is reflected in the broader venture capital interest in the AI security sector. As enterprises move beyond simple chatbots and begin to deploy agents that can take autonomous actions, the risks associated with prompt injection, hallucinated commands, and unintended data exposure have skyrocketed. Consequently, investors have been pouring significant capital into the space. Recent examples include Hadrian, which recently raised $40 million to expand its autonomous offensive security platform; Harmonic Security, which secured $50 million to focus on workforce AI protection; and Rein Security, which raised $25 million to guard AI agents at runtime.
The influx of capital into these firms signals a market-wide realization that the current security stack is insufficient for an era of agentic AI. As organizations continue to integrate these tools into their core workflows—from software development to customer support and data analysis—the ability to monitor and restrict agent activity has shifted from a "nice-to-have" feature to a fundamental requirement for enterprise compliance and risk management.
For Asymptote Labs, the challenge ahead lies in balancing the rapid innovation of the AI agent ecosystem with the rigorous, often conservative requirements of enterprise security departments. By anchoring their strategy in an open-source telemetry standard, the founders hope to build a neutral, foundational layer that becomes the industry standard for how enterprises observe and govern autonomous agents. As the company scales its engineering efforts, it intends to continue expanding support for new agent harnesses and frameworks, ensuring that as new AI development tools emerge, security teams remain one step ahead.
The company’s focus on runtime visibility is particularly timely. As AI agents move from experimental sandboxes to production environments, the traditional perimeter-based security model becomes increasingly obsolete. Agents operate within the heart of the network, and their actions are often indistinguishable from legitimate automated processes until a policy violation occurs. By focusing on the "runtime" aspect—the precise moment an agent attempts to manipulate a system—Asymptote aims to provide the granular control required to manage this new, highly autonomous workforce.
As Asymptote Labs moves forward with its expansion plans, the tech industry will be watching closely to see how the company’s open-source strategy impacts the wider security ecosystem. By fostering a standardized approach to agent telemetry, the startup is not only building a product but also attempting to shape the way the industry thinks about agent transparency. In an era where trust and accountability are the biggest hurdles to the widespread adoption of AI, the ability to provide verifiable, real-time control over autonomous systems may prove to be the most critical infrastructure development of the decade. With the support of its investors and a clear vision for the future of agent security, Asymptote Labs is positioning itself as a vital component in the enterprise AI security stack.
