Ultra Security raises $10M seed round to launch an AI agent security platform

Ultra Security has raised $10 million in seed funding and is launching a security platform for AI agents, according to a report from Pulse 2.0.

The headline ties the round directly to the product: the capital is meant to support the launch of an AI agent security platform. As companies hand more tasks to agents that call tools and touch data, securing agent identity, permissions, and behavior is emerging as a category of its own.

Unlike traditional application security, the risk with agents comes from autonomous action: they may hold API credentials, execute multi-step operations, and call external services without supervision. If a prompt injection or a misconfigured permission is exploited, the impact can travel along the tool chain.

A $10 million seed round is a mid-sized bet by current AI infrastructure standards — enough to fund product work and early customer validation, but not enough to prove the category has matured. The report summary available here does not detail investors or specific product capabilities, so outside assessment of the platform will have to wait for more disclosure.

The open question is whether the category can converge on reusable standards: agent permission models, audit log formats, runtime blocking policies. Vendors currently approach these differently, and whoever turns them into something enterprises will actually buy — rather than another logging dashboard — will shape how far the segment goes.

Next, watch whether Ultra Security publishes early customers, integrations with mainstream agent frameworks, and evidence of how well the platform blocks real attacks. For buyers, the value of this class of product ultimately comes down to how much controllable risk it removes.