Hush Security raises $30 million to govern enterprise AI agents

Why it matters

Hush is betting that identity infrastructure will become the enforcement layer for enterprise AI agents. Akamai's backing adds a potential distribution channel as autonomous software moves into production systems.

Micha Rave and his fellow Hush Security co-founders have raised a $30 million Series A to control how enterprise AI agents reach sensitive systems, SecurityWeek reported on July 28th.

Akamai Technologies joined the round as a strategic investor alongside returning backers Battery Ventures and YL Ventures. The financing brings Hush Security's total capital raised to $41 million, including an $11 million seed announced when Hush emerged from stealth in September 2025.

Hush plans to use the Series A to expand engineering and sales, with an emphasis on hiring in the United States. Hush also intends to add integrations with enterprise identity and access management systems and AI platforms, while building more corporate partnerships, according to Hush's funding announcement.

Rave, Hush's CEO, founded Hush with chief customer officer Chen Nisnkorn, chief technology officer Shmulik Ladkani and research and development chief Alon Horowitz. Rave has spent more than 25 years in cybersecurity and infrastructure technology, including product leadership roles at Proofpoint, Meta Networks, SanDisk and HARMAN. The four founders previously worked across Meta Networks and Proofpoint, giving Hush a founding group already familiar with selling identity infrastructure to large enterprises.

Meta Networks developed zero-trust network access technology before Proofpoint agreed to acquire it in 2019. The transaction was valued at $120 million, CTech reported. Hush applies a related identity-first approach to software processes that increasingly act without a human making each individual request.

From machine identities to agent governance

Hush emerged from stealth on September 10th, 2025, with a platform designed to replace static API keys, tokens and other long-lived credentials used by machines. Hush says its system instead grants narrowly scoped, just-in-time permissions when an application, workload or agent needs to reach another resource.

The Series A announcement extends that architecture into a broader governance layer for AI agents. Hush discovers agents operating inside an organization, maps the Model Context Protocol servers, tools and resources they use, and enrolls them in a central registry. Administrators can then assign policies, record actions and revoke an agent's access through a central kill switch.

"AI agents need strict identity, not just API keys," Rave said in the announcement.

The distinction matters because an agent can take a sequence of actions across databases, software tools and internal services after receiving an initial instruction. A conventional credential may establish that the agent can connect, while leaving security teams with less control over which actions it takes afterward. Hush is betting enterprises will require authorization and audit records at runtime, rather than relying on static permissions established before the agent begins working.

Hush says multiple Fortune 500 businesses use its platform. Hush also says IT infrastructure services provider Kyndryl has deployed the technology internally and begun reselling it to enterprise customers. Hush has not published revenue or annual recurring revenue figures alongside the round.

Akamai adds distribution weight

Akamai's participation gives Hush a strategic backer with existing relationships across enterprise networking, cloud infrastructure and cybersecurity. That position could help Hush reach security buyers already confronting the spread of service accounts, automated workloads and AI agents across several cloud environments.

The financing also places Hush in a growing group of security vendors treating non-human identity as a distinct control problem. Aembit focuses on workload identity and access, while Entro Security covers non-human identities and agentic AI. Vendors including Noma Security address adjacent AI discovery, risk and governance requirements.

Hush's pitch centers on enforcing permissions as an agent acts. That approach moves Hush beyond finding exposed credentials or cataloging agents and into the path of each access request, where reliability and integration depth become central buying criteria. The Series A will fund the engineering and partnerships needed to make that control layer work across the identity systems and AI platforms enterprises already use.