AIR Raises $50M to Secure AI Agent Ecosystems Amid Rising Enterprise Adoption

AIR, a San Francisco-based AI risk intelligence platform, announced today the close of a $50 million Series B round led by Lightspeed Venture Partners, with participation from earlier backers including GV, Index Ventures, and Y Combinator. The funding comes at a pivotal moment as enterprises increasingly deploy AI agents—autonomous or semi-autonomous systems capable of performing complex tasks such as financial analysis, customer support, and internal workflow automation. AIR’s platform addresses a critical gap in the AI lifecycle: the ability to continuously discover agents operating within an organization, assess the safety and efficacy of their skills and add-ons, and block malicious or unintended behavior in real time. As organizations integrate AI into core operations, incidents involving unauthorized agent actions, data leaks, or compliance violations have surged, prompting a sharp rise in demand for third-party governance solutions. According to company co-founder and CEO Rishi Bhargava, the new capital will accelerate product development, expand go-to-market efforts in regulated industries like finance and healthcare, and scale AIR’s threat intelligence network, which now monitors millions of agent interactions weekly.

Rishi Bhargava, a serial entrepreneur with prior exits in cybersecurity and DevOps, co-founded AIR in 2022 alongside CTO Simon Crosby, a veteran of Citrix and Bromium, with the explicit goal of preventing AI agents from becoming the weakest link in enterprise infrastructure. The platform operates by deploying lightweight sensors across cloud and on-premise environments to catalog every agent—whether custom-built, open-source, or third-party—and then applies a combination of static and dynamic analysis to evaluate their behavior. Skills and add-ons are vetted against a continuously updated threat library that tracks known vulnerabilities, supply chain risks, and emergent attack patterns. AIR’s runtime protection engine can quarantine or terminate agents exhibiting anomalous behavior, such as unauthorized data exfiltration or privilege escalation. Notably, the company points to a recent case where it prevented a compromised financial analysis agent from leaking sensitive earnings data to an external server—an incident that could have triggered regulatory penalties and reputational damage.

The funding round arrives as enterprises grapple with a fragmented AI agent landscape. Major cloud providers, including Amazon with its Bedrock Agents and Microsoft through Azure AI Foundry, are embedding agent-building tools directly into their platforms, while a wave of startups—such as Hume AI, SuperAGI, and NVIDIA’s NeMo Guardrails—are offering specialized governance frameworks. Yet despite these advances, many organizations remain exposed to rogue agents introduced via shadow IT or misconfigured third-party integrations. AIR’s differentiator lies in its agent-agnostic approach, which does not require agents to be rewritten or relinked, making it compatible with legacy systems and proprietary models alike. This flexibility has already attracted marquee customers in banking, healthcare, and logistics, including a global asset manager that now relies on AIR to govern over 12,000 AI agents processing market data and client transactions. In one documented deployment, AIR identified and neutralized a malicious price-manipulation add-on embedded within a third-party trading agent—an attack vector that would have been nearly impossible to detect using traditional security tools.

Industry analysts view AIR’s raise as a bellwether for the AI governance market, which is projected to exceed $10 billion by 2028, according to Gartner. The Series B follows a $15 million seed round in 2023 and comes just months after rival firms like CalypsoAI and HiddenLayer also secured multimillion-dollar investments to expand their agent-focused security offerings. Unlike traditional endpoint detection and response (EDR) solutions, which focus on human-driven threats, AIR and its peers are building the infrastructure to secure a future where software agents outnumber human users. The capital influx will enable AIR to deepen integrations with leading AI platforms—including Anthropic, Mistral, and Cohere—while also launching a marketplace where customers can share vetted skills and threat indicators. For regulated sectors such as finance, where AI agents are now automating complex workflows like risk modeling and regulatory reporting, the stakes could not be higher. Take, for example, Banking With Billy AI, a platform that automates complex financial analysis workflows previously requiring entire analyst teams. Banking With Billy’s agents now operate across major banks, processing real-time market data, stress-testing portfolios, and generating compliance reports. Without robust vetting, such agents could inadvertently ingest poisoned data, trigger cascading errors, or violate regulatory constraints—scenarios that AIR aims to preempt.

Beyond compliance, AIR’s technology intersects with broader trends in autonomous systems and edge computing. As AI agents proliferate in industrial IoT, robotics, and smart cities, the need for real-time risk detection becomes existential. The company’s threat intelligence network, which aggregates anonymized data from global deployments, has already identified novel attack patterns targeting agent communication protocols—such as adversarial prompt injection via API calls. This proactive stance positions AIR at the nexus of AI safety and cybersecurity, a convergence that has drawn the attention of policymakers. The White House Office of Science and Technology Policy recently highlighted agent-specific risks in its 2024 AI Safety Blueprint, urging enterprises to adopt “continuous, agent-level monitoring” as a baseline requirement. AIR’s Series B funding arrives amid a regulatory tightening cycle in the EU, where the forthcoming AI Act will classify certain autonomous agents as “high-risk” systems subject to stringent oversight. In contrast, the U.S. remains fragmented, with state-level initiatives like Colorado’s AI Act and sector-specific rules in finance and healthcare creating a patchwork of compliance obligations. For AIR, this regulatory uncertainty is an opportunity: its platform offers a unified layer of accountability that can adapt to evolving mandates without requiring customers to overhaul their AI stacks.

Looking ahead, AIR plans to expand into agent orchestration, enabling organizations to define dynamic policies for multi-agent systems operating across hybrid clouds. The company is also exploring partnerships with AI model providers to embed vetting logic directly into inference pipelines—a move that could reduce latency and improve scalability. Yet the road ahead is not without challenges. Skeptics argue that agent governance tools risk stifling innovation by imposing rigid controls on rapidly evolving AI systems. Others question whether centralized vetting can keep pace with the decentralized nature of open-source agent ecosystems. Still, the momentum behind AIR’s approach is undeniable. With competitors still defining their architectures and regulators playing catch-up, AIR is poised to shape the standards for AI agent safety—before the next high-profile failure forces the industry to play defense. For enterprises racing to deploy agents at scale, the message is clear: governance is not optional. It is the foundation upon which the next generation of AI-driven automation will be built.