Keycard Recognized as AI Agent Security Leader as CRN Names It a Top Cloud Startup to Watch in 2026

New updates have been reported about Keycard.

Keycard, a San Francisco–based provider of identity and access infrastructure for AI agents, has been named one of CRN’s 10 cloud computing startups to watch in 2026, underscoring growing market recognition of its role in securing autonomous AI workflows. The accolade follows CRN’s 2025 inclusion of Keycard among the 10 “coolest” cloud startups, and comes on the heels of the company’s October 2025 public launch, backed by $38 million in seed and Series A funding from Andreessen Horowitz, boldstart ventures, Acrew Capital and others. CRN highlighted Keycard’s positioning as a “missing ingredient” for secure, trusted agent access, emphasizing its ability to connect AI agents and applications to a managed enterprise cloud with optional private networking, and to broker agent-to-service connections across clouds and teams.

Keycard’s platform is designed to address a central concern for enterprises deploying AI agents at scale: how to grant autonomy without creating uncontrolled access risk. The system assigns identity to AI agents, applies task-based permissions, enforces policy in real time, and logs all activity, using ephemeral, identity-bound tokens, mixed delegation chains, and task-scoped policy to tightly constrain what agents can do and for how long. The company extended its technical footprint in November 2025 with the acquisition of Runebook, broadening its SDK ecosystem so enterprises can build and adopt MCP-powered AI agents and tools without deep in-house identity expertise or exposure to unconstrained access. CEO and co-founder Ian Livingstone said the company’s model is built on “access that knows who’s asking and why, scoped to the task and gone when it’s done,” positioning Keycard as a foundational security layer for the shift from static, human-driven workflows to machine-driven, autonomous applications. For executives, the developments signal that Keycard is emerging as a key infrastructure provider for governing AI agents in production environments, with growing third-party validation and a platform directly aimed at security, compliance, and operational risk management in enterprise AI adoption.